CN113255102A - Method and device for predicting carbon content and temperature of molten steel at converter end point - Google Patents
Method and device for predicting carbon content and temperature of molten steel at converter end point Download PDFInfo
- Publication number
- CN113255102A CN113255102A CN202110444384.6A CN202110444384A CN113255102A CN 113255102 A CN113255102 A CN 113255102A CN 202110444384 A CN202110444384 A CN 202110444384A CN 113255102 A CN113255102 A CN 113255102A
- Authority
- CN
- China
- Prior art keywords
- cases
- case
- converter
- temperature
- end point
- 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
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
-
- C—CHEMISTRY; METALLURGY
- C21—METALLURGY OF IRON
- C21C—PROCESSING OF PIG-IRON, e.g. REFINING, MANUFACTURE OF WROUGHT-IRON OR STEEL; TREATMENT IN MOLTEN STATE OF FERROUS ALLOYS
- C21C5/00—Manufacture of carbon-steel, e.g. plain mild steel, medium carbon steel or cast steel or stainless steel
- C21C5/28—Manufacture of steel in the converter
- C21C5/30—Regulating or controlling the blowing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Abstract
The invention discloses a method and a device for predicting carbon content and temperature of molten steel at a converter end point, wherein the method comprises the following steps: taking each converter smelting process as a case respectively, and describing the converter smelting processes by case; the case description comprises case characteristic description and solution description of the case, and the case characteristics comprise single-value type influence factors and time sequence type influence factors; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter; taking the current converter smelting process as a problem case and the historical converter smelting process as a historical case, and retrieving similar cases in the historical cases based on a case reasoning algorithm; and based on the retrieved similar cases and the carbon content and the temperature of the molten steel at the converter end point, obtaining the carbon content and the temperature predicted value of the molten steel at the converter end point of the problem case through reusing the cases. The prediction accuracy of the method is superior to that of the existing prediction model, and the method can meet the requirements of converter field production.
Description
Technical Field
The invention relates to the technical field of converter process end point control, in particular to a method and a device for predicting carbon content and temperature of molten steel at a converter end point.
Background
Converter steelmaking is a very complex multi-element multi-phase high-temperature physicochemical process, and has the remarkable characteristics of high reaction speed, a plurality of influencing factors and complex reaction. The converter endpoint control mainly refers to the carbon content and temperature of the endpoint. Inaccurate end point control can result in a series of hazards such as increase of oxygen content of molten steel, increase of iron loss, prolongation of blowing time, reduction of service life of furnace lining and the like. Therefore, the improvement of the hit rate of the converter end point control has important significance for improving the product quality, accelerating the production rhythm and improving the enterprise profit. The carbon content and the temperature of the molten steel at the end point of the converter are one of indexes for controlling the end point of the converter process, and an accurate end point prediction model is favorable for improving the hit rate of the end point of the converter.
The current end point control model of the converter can be divided into a static control model and a dynamic control model, wherein the static control model is the basis of the dynamic control model and can be divided into a mechanism model and a data driving model according to a modeling principle. Because the mechanism model is relatively ideal, parameters in the model cannot be obtained due to field condition limitation, and the precision of the model is low. With the rapid development of automation and informatization of steel mills, each large steel mill establishes a large data platform and collects a large amount of production data, and a solution is provided for improving the hit rate of the converter end point based on a data-driven end point prediction model.
At present, a plurality of methods for predicting the end point of each process are provided, and algorithms such as a support vector regression machine, a neural network, a decision tree and case reasoning are commonly used. The high break et al propose a converter static prediction model based on an improved twin support vector machine. Korean and Min et al establish a converter steelmaking dynamic control model based on ANFIS and a robust correlation vector machine. He Fei et al established a converter endpoint phosphorus content prediction model based on PCA and BP neural networks. Luwu et al established an LF endpoint temperature prediction model based on a limit learning machine. The field wisdom et al establishes an extreme learning machine model based on an improved AdaBoost. Korea-sensitive et al, based on the model for predicting the steelmaking end point of an oxygen converter by using a membrane algorithm evolution extreme learning machine. The Wangxiang et al respectively establish LF endpoint temperature prediction models based on random forests and integrated regression trees based on bootstrap feature subsets. The He Fei et al propose an LF endpoint prediction model based on case-based reasoning. Von Kai et al propose a case reasoning model based on mechanism model similarity to predict the final phosphorus content of a special dephosphorization converter, and Wangxi et al propose a converter steelmaking static control model based on a CBR model of causal relationship. In addition, Jiang Shenglong et al proposed a mixed model based on multiple linear regression and Gaussian process regression to predict converter oxygen consumption. Severe waves et al established a prediction model of endpoint carbon content based on a genetic algorithm's kernel partial least squares regression (GA-KPLSR) method. The process et al propose a data-driven multi-task learning (MTL) steelmaking endpoint prediction method.
The above prediction models do not fully consider time sequence type process parameters, such as lance position change, oxygen supply flow, bottom blowing gas flow and the like in the converter blowing process, and these time sequence type process parameters have great influence on the converter terminal point composition and temperature, so the existing prediction models are not ideal enough for the prediction of the converter terminal point.
Disclosure of Invention
The invention provides a method and a device for predicting carbon content and temperature of molten steel at a converter end point, which are used for solving the technical problem that the prediction effect of the conventional prediction model on parameters such as carbon content, temperature and the like of the molten steel at the converter end point is not ideal due to the fact that the conventional prediction model does not consider the process parameters of time sequence types in converter procedures.
In order to solve the technical problems, the invention provides the following technical scheme:
in one aspect, the invention provides a method for predicting carbon content and temperature of molten steel at a converter endpoint, which comprises the following steps:
taking each converter smelting process as a case respectively, and describing the converter smelting processes by case; the case description comprises case characteristic description and case solution description, and the case characteristics comprise single-value type influence factors and time sequence type influence factors which influence the carbon content and the temperature of molten steel at the converter end point; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter in the corresponding converter smelting process;
searching similar cases with similarity between the historical cases and the problem cases meeting preset requirements in the historical cases and the converter end point molten steel carbon content and temperature of the similar cases based on a case reasoning algorithm according to case characteristic description of the problem cases by taking the current converter smelting process as the problem cases and the historical converter smelting process as the historical cases;
and obtaining a predicted value of the carbon content and the temperature of the converter end point molten steel corresponding to the problem case through reusing the cases based on the retrieved similar cases and the carbon content and the temperature of the converter end point molten steel corresponding to the similar cases.
Further, the single-value type influence factors comprise molten iron information, auxiliary raw material adding amount and gas consumption; wherein the content of the first and second substances,
the molten iron information comprises molten iron temperature, molten iron weight and molten iron composition; the molten iron components comprise molten iron carbon content, molten iron silicon content, molten iron manganese content, molten iron phosphorus content and molten iron sulfur content;
the auxiliary raw materials comprise scrap steel, lime, light-burned dolomite and sinter;
the gas consumption comprises oxygen consumption and argon consumption;
the time sequence type influencing factors comprise oxygen supply flow, oxygen lance position and bottom argon blowing flow.
Further, the step of retrieving a similar case, in which the similarity between the historical case and the problem case meets preset requirements, based on a case reasoning algorithm according to the case feature description of the problem case includes:
calculating the similarity of the single-value type influence factors of the historical cases and the problem cases;
calculating the similarity of time sequence type influence factors of the historical cases and the problem cases;
carrying out weighted combination on the similarity of the single-value type influence factors and the similarity of the sequence type influence factors of the historical cases and the problem cases to obtain the comprehensive similarity between the corresponding historical cases and the problem cases;
and screening out similar cases meeting preset requirements from the historical cases according to the comprehensive similarity.
Further, the calculating of the similarity of the single-value type influence factors of the historical cases and the problem cases comprises calculating the similarity of the single-value type influence factors of the historical cases and the problem cases based on Euclidean distance.
Further, the calculating the similarity of the time sequence type influence factors of the historical cases and the problem cases comprises: and calculating the similarity of the time sequence type influence factors of the historical cases and the problem cases based on a dynamic time warping algorithm.
Further, the obtaining of the predicted values of the carbon content and the temperature of the converter end point molten steel corresponding to the problem case by reusing the cases based on the retrieved similar cases and the converter end point molten steel corresponding to the similar cases includes:
and solving by adopting a k nearest neighbor method based on the retrieved similar cases and the carbon content and the temperature of the molten steel at the converter end point corresponding to the similar cases to obtain a predicted value of the carbon content and the temperature of the molten steel at the converter end point corresponding to the problem case.
In another aspect, the present invention further provides a device for predicting carbon content and temperature of molten steel at a converter end point, comprising:
the case description module is used for respectively regarding each converter smelting process as a case and describing the converter smelting processes; the case description comprises case characteristic description and case solution description, and the case characteristics comprise single-value type influence factors and time sequence type influence factors which influence the carbon content and the temperature of molten steel at the converter end point; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter in the corresponding converter smelting process;
the case retrieval module is used for retrieving similar cases and converter end point molten steel carbon contents and temperatures of the similar cases, wherein the similarity between the similar cases and the problem cases in the historical cases meets preset requirements, and the converter end point molten steel carbon contents and temperatures of the similar cases are obtained based on case characteristic description of the case description module on the problem cases by taking the current converter smelting process as the problem cases and the historical converter smelting process as the historical cases;
and the case reuse module is used for obtaining the predicted values of the carbon content and the temperature of the converter end point molten steel corresponding to the problem case through case reuse based on the similar cases searched by the case search module and the converter end point molten steel corresponding to the similar cases.
Further, the single-value type influence factors comprise molten iron information, auxiliary raw material adding amount and gas consumption; wherein the content of the first and second substances,
the molten iron information comprises molten iron temperature, molten iron weight and molten iron composition; the molten iron components comprise molten iron carbon content, molten iron silicon content, molten iron manganese content, molten iron phosphorus content and molten iron sulfur content;
the auxiliary raw materials comprise scrap steel, lime, light-burned dolomite and sinter;
the gas consumption comprises oxygen consumption and argon consumption;
the time sequence type influencing factors comprise oxygen supply flow, oxygen lance position and bottom argon blowing flow.
Further, the case retrieval module is specifically configured to:
calculating the similarity of the single-value type influence factors of the historical cases and the problem cases;
calculating the similarity of time sequence type influence factors of the historical cases and the problem cases;
carrying out weighted combination on the similarity of the single-value type influence factors and the similarity of the sequence type influence factors of the historical cases and the problem cases to obtain the comprehensive similarity between the corresponding historical cases and the problem cases;
according to the comprehensive similarity, screening out similar cases meeting preset requirements from the historical cases;
the calculating the similarity of the single-value type influence factors of the historical cases and the problem cases comprises the following steps: calculating the similarity of single-value type influence factors of the historical case and the problem case based on the Euclidean distance; the calculating the similarity of the time sequence type influence factors of the historical cases and the problem cases comprises the following steps: and calculating the similarity of the time sequence type influence factors of the historical cases and the problem cases based on a dynamic time warping algorithm.
Further, the case reuse module is specifically configured to:
and solving by adopting a k nearest neighbor method based on the retrieved similar cases and the carbon content and the temperature of the molten steel at the converter end point corresponding to the similar cases to obtain a predicted value of the carbon content and the temperature of the molten steel at the converter end point corresponding to the problem case.
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:
in the invention, each converter smelting process is respectively regarded as a case, and the converter smelting processes are described by case; the case description comprises case characteristic description and solution description of the case, and the case characteristics comprise single-value type influence factors and time sequence type influence factors; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter; taking the current converter smelting process as a problem case and the historical converter smelting process as a historical case, and retrieving similar cases in the historical cases based on a case reasoning algorithm; and based on the retrieved similar cases and the carbon content and the temperature of the molten steel at the converter end point, obtaining the carbon content and the temperature predicted value of the molten steel at the converter end point of the problem case through reusing the cases. The final prediction precision is superior to that of the existing prediction model, and the requirement of converter field production can be met.
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 flow chart of a method for predicting carbon content and temperature of molten steel at a converter end point according to an embodiment of the present invention;
fig. 2 is a schematic diagram of case characterization provided by an embodiment of the present invention;
FIG. 3 shows the difference wtimeSetting an evaluation index change schematic diagram of a next carbon content prediction result;
FIG. 4 shows the difference wtimeAnd setting an evaluation index change schematic diagram of the lower temperature prediction result.
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 method for predicting carbon content and temperature of molten steel at the end point of a converter, and provides a case reasoning model (CBR _ TM) based on time sequence data to predict the carbon content and temperature at the end point of the converter according to the type of process parameters in the smelting process of the converter. The characteristics of the cases in the model not only comprise single-value type influence factors (molten iron components, weight, temperature and the like) which influence the end point carbon content and the temperature, but also comprise time sequence type influence factors such as gun position change, oxygen supply flow and the like, and the single-value type influence factor similarity and the time sequence type influence factor similarity between the cases are respectively calculated based on Euclidean distance and a dynamic time warping algorithm in the case retrieval process, and are weighted and combined to obtain comprehensive similarity. The method may be implemented by an electronic device, which may be a terminal or a server. The execution flow of the method is shown in fig. 1, and comprises the following steps:
s101, regarding each converter smelting process as a case, and describing the converter smelting processes in case; the case description comprises case characteristic description and case solution description, and the case characteristics comprise single-value type influence factors and time sequence type influence factors which influence the carbon content and the temperature of molten steel at the converter end point; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter in the corresponding converter smelting process;
s102, searching similar cases with the similarity between the historical cases and the problem cases meeting preset requirements in the historical cases and the converter end point molten steel carbon content and temperature of the similar cases by taking the current converter smelting process as the problem cases and the historical converter smelting process as the historical cases and based on case reasoning algorithm according to case characteristic description;
s103, based on the retrieved similar cases and the carbon content and the temperature of the molten steel at the converter end point corresponding to the similar cases, the predicted values of the carbon content and the temperature of the molten steel at the converter end point corresponding to the problem cases are obtained through reuse of the cases.
Specifically, as shown in fig. 2, the present embodiment divides the influencing factors into two types of data, a single-value type influencing factor and a time-series type influencing factor, according to the data types of the influencing factors of the converter endpoint composition and the temperature.
The single value type influence factors mainly comprise molten iron information, auxiliary raw material adding amount, gas consumption and the like; the molten iron information mainly comprises molten iron temperature, molten iron weight, molten iron components and the like; the molten iron mainly comprises the components of molten iron, such as the carbon content, the silicon content, the manganese content, the phosphorus content, the sulfur content and the like of the molten iron; the addition of the auxiliary raw materials mainly comprises the addition of scrap steel, the addition of lime, the addition of light-burned dolomite, the addition of sinter and the like; the gas consumption mainly includes oxygen consumption, argon consumption, and the like.
The time sequence type influencing factors mainly comprise oxygen supply flow, oxygen lance position, bottom argon blowing flow and the like.
Thus a case can be described as: the case is { single value data set, time series data set }, wherein, single value data set is { molten iron information (molten iron composition, temperature, weight, etc.), auxiliary raw materials (scrap steel, lime, boiled dolomite, etc.), gas consumption (oxygen and argon) }, time series data set is { nutrient supply flow, oxygen lance position and argon flow }.
Based on the above, the case retrieval process of this embodiment includes:
calculating the similarity of the single-value type influence factors of the historical cases and the problem cases;
calculating the similarity of time sequence type influence factors of the historical cases and the problem cases;
carrying out weighted combination on the similarity of the single-value type influence factors and the similarity of the sequence type influence factors of the historical cases and the problem cases to obtain the comprehensive similarity between the corresponding historical cases and the problem cases;
and screening out similar cases meeting preset requirements from the historical cases according to the comprehensive similarity. Here, the cases may be sorted in descending order according to the magnitude of the integrated similarity, and then the first few cases may be selected.
The similarity calculation methods of the single-value type influence factors include Euclidean distance similarity, gray distance similarity and the like. The present embodiment adopts the euclidean distance similarity.
Assuming that the number of influencing factors of a case is n, the jth factor of the historical case in the case base is xjIn the case of the problem, the j-th influencing factor is yjAnd the weight of the jth factor is wjThen, the euclidean distance formula between the problem case and the historical case in the case base is shown in equations (1) - (2):
in the formula: m is the number of influencing factors of the case; x is the number ofjIs the jth influencing factor of the problem case; y isjThe jth influence factor of the historical case in the case base; w is ajIs the weight of the jth influencing factor.
The single-value data similarity between the history case and the problem case is as follows:
method for measuring similarity of time sequence type influence factorsThere are also many, and can be divided into two categories: the method comprises a similarity measurement method based on track points and a similarity measurement method based on track segments. The similarity measurement method based on the track points can be divided into a global matching measurement method and a local matching measurement method. The global matching metric method includes an euclidean distance method, a Dynamic Time Warping (DTW) method, and an edit distance method (ERP), and in consideration of the characteristic that time series data in the converter have different lengths, the dynamic time warping method is used in the embodiment to calculate the similarity of the time series data. Assume that two time series a ═ a1,A2,...,Ai,...,An}，B＝{B1,B2,...,Bi,...,BmDTW obtains the minimum distance between 2 time sequences by bending the time axis, determines the best matching relation of each point, and A matched with each otheriAnd BjThe difference between them is the distance at that moment.
To determine the best matching relationship, a, B form an n × m DTW matrix d:
in the DTW matrix d, the basic idea of dynamic programming is applied from the starting point (1,1) to the end point (N, M), using the formula:
Dn,m＝dn,m+min{Dn-1,m,Dn-1,m-1,Dn,m-1} (4)
in the formula: dn,mAnd calculating the local optimal cumulative distance according to the cumulative distance between the current point and the previous point.
Defining the time sequence similarity of the time sequences A and B as follows:
after the single-value data similarity and the time sequence data similarity between the historical cases and the problem cases are calculated by the method, the single-value data similarity and the time sequence data similarity can be weighted and summed to obtain the comprehensive similarity of the cases, and the formula is as follows:
in the formula: n is the number of time series data variables; w is asingle,wtimeWeights, w, for single-valued data similarity and time-series data similarity, respectivelysingle+wtimeSpecific values are determined after experimental optimization, namely 1.
Further, after completing the case search, the case reuse process of this embodiment is as follows:
the k-nearest neighbor method is adopted for solving, and the calculation formula is shown as formula (7):
in the formula: k is the number of case reuse, SiAs the integrated similarity of cases, TiThe solution for a similar case, namely the carbon content and the temperature at the end of the converter.
Next, the accuracy of the prediction model constructed by the method of the present embodiment is verified by using an application example.
1 data set
In this embodiment, 946 actual production data of the SHPC steel converter collected by the B steel mill are used for verification, wherein 846 data are used as a training set, and 100 data are used as a test set. And selecting 16 influence factors according to field data, wherein 13 single value types are respectively molten iron temperature, molten iron weight, molten iron carbon content, molten iron silicon content, molten iron manganese content, molten iron phosphorus content, scrap steel amount, lime addition amount, dolomite addition amount, sinter ore addition amount, heat supply agent, total oxygen consumption and total argon consumption, and 3 time sequence data are respectively oxygen supply flow, oxygen lance position and bottom blowing argon flow. The specific influencing factor statistics are shown in tables 1 and 2.
TABLE 1 statistical results of the influencing factors (single-valued type) of the converter end point carbon temperature
| Influencing factor | Numbering | Maximum value | Minimum value | Mean value | Standard deviation of |
| Temperature of molten iron/. degree.C | X1 | 1437 | 1080 | 1264.22 | 120.40 |
| Weight of molten iron/t | X2 | 297 | 220 | 275.53 | 10.55 |
| Carbon content of molten iron/%) | X3 | 4.6978 | 3.8001 | 4.2925 | 0.1442 |
| Silicon content of molten iron% | X4 | 0.50622 | 0.00473 | 0.15441 | 0.09165 |
| Iron manganese content/%) | X5 | 0.27224 | 0.00873 | 0.16096 | 0.02762 |
| Phosphorus content of molten iron% | X6 | 0.13164 | 0.04730 | 0.10261 | 0.01130 |
| Amount of scrap steel/t | X7 | 58.9 | 24.3 | 45.22 | 4.453 |
| Lime addition/t | X8 | 16.781 | 1.113 | 10.360 | 1.774 |
| Dolomite addition/t | X9 | 15.121 | 2.311 | 4.042 | 1.009 |
| Sinter addition/t | X10 | 11.803 | 0 | 2.6948 | 2.2516 |
| Heat-supplementing agent/t | X11 | 4.078 | 0 | 0.5964 | 0.82364 |
| Total oxygen consumption/Nm 3 | X12 | 16990 | 11700 | 14951 | 631 |
| Total argon consumption/Nm 3 | X13 | 123 | 10 | 40.77 | 34.29 |
| TSO[C]/％ | Y1 | 0.1281 | 0.01848 | 0.0443 | 0.01546 |
| TSO[T]/℃ | Y2 | 1715 | 1620 | 1675 | 18.2 |
TABLE 2 statistical results of the factors (time series type) affecting the end point carbon temperature of the converter
2 evaluation index
To evaluate the effectiveness of the prediction model, three indicators were used for evaluation, Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and hit rate (HitRate). The calculation formula is as follows:
wherein: y isiIs the actual value for the i-th case,
the predicted value of the ith case is, and n is the number of cases in the test set. Errorband is the error range, specifically, in this example, the error ranges of the prediction of the carbon content and the temperature are + -0.02% and + -15 deg.C, respectively.
3 results and analysis
The parameters of the time-series data-based improved CBR model (CBR _ TM) provided by the present embodiment are set as follows: the data standardization of the single-value type data adopts (-1,1) standardization, the similarity calculation method is Euclidean distance, the weight calculation method is an entropy weight method, and the reuse number of cases is 3.
The single-value type data weight calculation results are as follows:
TABLE 3 weight of single value data
| X1 | X2 | X3 | X4 | X5 | X6 | X7 | |
| Weight | 0.0134 | 0.0119 | 0.0099 | 0.0641 | 0.0116 | 0.0082 | 0.0578 |
| X8 | X9 | X10 | X11 | X12 | X13 | ||
| Weight | 0.0162 | 0.0391 | 0.1653 | 0.4740 | 0.0111 | 0.1174 |
And the time sequence data similarity calculation adopts a dynamic warping (DTW) algorithm, and the data normalization also adopts (-1,1) normalization. w is asingle,wtimeIs an important parameter in the model, and the present example studies wtimeInfluence on the prediction accuracy of the model, in particular wtimeSet up as shown in Table 4, where wtimeWhen 0, only single value type is considered for the modelData, wtimeWhen 1, only time series type data is considered for the model.
Table 4 model wsingle,wtimeIs provided with
| Experiment number | Model numbering | wsingle | wtime |
| 1 | CBR_TM(1,0) | 1 | 0 |
| 2 | CBR_TM(0.9,0.1) | 0.9 | 0.1 |
| 3 | CBR_TM(0.8,0.2) | 0.8 | 0.2 |
| 4 | CBR_TM(0.7,0.3) | 0.7 | 0.3 |
| 5 | CBR_TM(0.6,0.4) | 0.6 | 0.4 |
| 6 | CBR_TM(0.5,0.5) | 0.5 | 0.5 |
| 7 | CBR_TM(0.4,0.6) | 0.4 | 0.6 |
| 8 | CBR_TM(0.3,0.7) | 0.3 | 0.7 |
| 9 | CBR_TM(0.2,0.8) | 0.2 | 0.8 |
| 10 | CBR_TM(0.1,0.9) | 0.1 | 0.9 |
| 11 | CBR_TM(0,1) | 0 | 1 |
By making statistics of differenceswtimeThe set model terminal temperature and carbon content prediction results can be used for calculating different wtimeThe evaluation index statistics of the lower prediction model are shown in fig. 3 and 4, and it can be seen from the figures that as w is increasedtimeThe MAE and the RMSE in the evaluation indexes of carbon content and temperature prediction show the trend of firstly decreasing and then increasing, and the index Hitrate shows the trend of firstly increasing and then decreasing, namely, the model prediction precision is firstly increased and then decreased. For carbon content prediction, at wtimeThe prediction accuracy is highest when the value is 0.4, and the MAE, RMSE and Hitrate of the model are respectively 6.034 per mill, 7.032 per mill and 89 percent, and w is equal totimeThe MAE and RMSE indices decreased by 1.51 and 1.677, respectively, and the HitRate index increased by 9% compared to 0, for the temperature prediction, also at wtimeThe prediction accuracy is highest when the value is 0.4, and the MAE, RMSE and Hitrate of the model are respectively 6.034 per mill, 7.032 per mill and 89 percent, and w is equal totimeThe MAE and RMSE indices decreased by 1.048 and 1.310 respectively and the HitRate index increased by 9% compared to 0.
Further analysis shows that the input of the prediction model simultaneously comprises single-value type data and time sequence type data, which is helpful for improving the accuracy of the prediction model, but the respective weights are not easy to be too high, and if the influence of a certain type on the end point control is neglected, the accuracy of the model prediction is reduced.
In addition, in order to further verify the accuracy of the prediction model provided by the embodiment, the embodiment also establishes the prediction accuracy based on the SVR and BPNN models, and only single-subtype data is taken as input in the models.
The SVR model is constructed by calling an SVR algorithm in a python data mining toolkit scimit-spare, a polynomial core (poly core) is selected by a core function in the parameters of the model, and the frequency of the polynomial core function is 3.
The BPNN model is constructed by calling a python deep learning toolkit tensorflow, the network structure of the model parameter is 4 layers, the nodes of an input layer are 16, the number of hidden layers is 2 layers, the number of the nodes of a first hidden layer is 8, the number of the hidden layers of a second layer is 5, the number of the nodes of an output layer is 1, and the activation function is relu.
The prediction results of each model are counted to obtainCBR_TM(0.6,0.4)The model is lower than the SVR and BPNN models in both MAE and RMSE indexes, and the model prediction hit rate is higher than the SVR and BPNN models.
In summary, the embodiment provides a method for predicting carbon content and temperature of molten steel at the end point of a converter, and a case-based reasoning model (CBR _ TM) based on time sequence data is established to predict the carbon content and temperature of the molten steel at the end point of the converter. The input of the model comprises single-value type data such as molten iron components, temperature and the like, and also comprises time sequence type data such as gun position change, oxygen supply flow and the like, so that case characteristics are more comprehensive, a similarity calculation method of the time sequence type data is provided based on a dynamic time warping algorithm, and the accuracy of case retrieval is improved. Moreover, the present embodiment also studies the weighting factor w based on the field production datatimeThe influence on the accuracy of the prediction model is shown by experimental results, and the prediction accuracy of the carbon content and the temperature is dependent on wtimeShows a trend of increasing first and then decreasing, and is at wtimeThe highest prediction accuracy is achieved when the carbon content is 0.4, and the carbon content is predicted at wtimeThe MAE, RMSE and HitRate for the model 0.4 were 6.034%, 7.032%, and 89%, respectively, vs. wtimeThe MAE and RMSE indices decreased by 1.51 and 1.677, respectively, and the HitRate index increased by 9% compared to 0, for the temperature prediction, also at wtimeThe MAE, RMSE and HitRate for the model at 0.4 were 8.361%, 9.687% and 89%, respectively, vs wtimeThe MAE and RMSE indices decreased by 1.048 and 1.310 respectively and the HitRate index increased by 9% compared to 0. In addition, the embodiment further compares the established prediction model with the indexes based on the SVR and the BPNN model, and the result shows that both the MAE and RMSE indexes of the prediction model are lower than those of the SVR and the BPNN model, so that the effectiveness of the prediction model is proved, and the prediction accuracy of the prediction model can meet the field requirements.
Second embodiment
The embodiment provides a device for predicting carbon content and temperature of molten steel at the end point of a converter, which comprises:
the case description module is used for respectively regarding each converter smelting process as a case and describing the converter smelting processes; the case description comprises case characteristic description and case solution description, and the case characteristics comprise single-value type influence factors and time sequence type influence factors which influence the carbon content and the temperature of molten steel at the converter end point; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter in the corresponding converter smelting process;
the case retrieval module is used for retrieving similar cases and converter end point molten steel carbon contents and temperatures of the similar cases, wherein the similarity between the similar cases and the problem cases in the historical cases meets preset requirements, and the converter end point molten steel carbon contents and temperatures of the similar cases are obtained based on case characteristic description of the case description module on the problem cases by taking the current converter smelting process as the problem cases and the historical converter smelting process as the historical cases;
and the case reuse module is used for obtaining the predicted values of the carbon content and the temperature of the converter end point molten steel corresponding to the problem case through case reuse based on the similar cases searched by the case search module and the converter end point molten steel corresponding to the similar cases.
The converter end point molten steel carbon content and temperature prediction device of the present embodiment corresponds to the converter end point molten steel carbon content and temperature prediction method of the first embodiment described above; the functions realized by the functional modules in the converter end point molten steel carbon content and temperature prediction device of the embodiment correspond to the flow steps in the converter end point molten steel carbon content and temperature prediction method 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 method for predicting carbon content and temperature of molten steel at a converter endpoint is characterized by comprising the following steps:
taking each converter smelting process as a case respectively, and describing the converter smelting processes by case; the case description comprises case characteristic description and case solution description, and the case characteristics comprise single-value type influence factors and time sequence type influence factors which influence the carbon content and the temperature of molten steel at the converter end point; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter in the corresponding converter smelting process;
searching similar cases with similarity between the historical cases and the problem cases meeting preset requirements in the historical cases and the converter end point molten steel carbon content and temperature of the similar cases based on a case reasoning algorithm according to case characteristic description of the problem cases by taking the current converter smelting process as the problem cases and the historical converter smelting process as the historical cases;
and obtaining a predicted value of the carbon content and the temperature of the converter end point molten steel corresponding to the problem case through reusing the cases based on the retrieved similar cases and the carbon content and the temperature of the converter end point molten steel corresponding to the similar cases.
2. The method of predicting carbon content and temperature of molten steel at a converter end point of claim 1, wherein the single value type influencing factors include molten iron information, an addition amount of auxiliary materials, and a gas consumption amount; wherein the content of the first and second substances,
the molten iron information comprises molten iron temperature, molten iron weight and molten iron composition; the molten iron components comprise molten iron carbon content, molten iron silicon content, molten iron manganese content, molten iron phosphorus content and molten iron sulfur content;
the auxiliary raw materials comprise scrap steel, lime, light-burned dolomite and sinter;
the gas consumption comprises oxygen consumption and argon consumption;
the time sequence type influencing factors comprise oxygen supply flow, oxygen lance position and bottom argon blowing flow.
3. The method for predicting the carbon content and the temperature of the molten steel at the converter endpoint according to claim 1, wherein the step of retrieving similar cases, the similarity of which with the problem cases meets preset requirements, in the historical cases according to the case characteristic description of the problem cases and based on a case reasoning algorithm comprises the following steps:
calculating the similarity of the single-value type influence factors of the historical cases and the problem cases;
calculating the similarity of time sequence type influence factors of the historical cases and the problem cases;
carrying out weighted combination on the similarity of the single-value type influence factors and the similarity of the sequence type influence factors of the historical cases and the problem cases to obtain the comprehensive similarity between the corresponding historical cases and the problem cases;
and screening out similar cases meeting preset requirements from the historical cases according to the comprehensive similarity.
4. The method for predicting the carbon content and the temperature of the molten steel at the end point of the converter according to claim 3, wherein the calculating the similarity of the single-value type influence factors of the historical case and the problem case comprises: and calculating the similarity of the single-value type influence factors of the historical case and the problem case based on the Euclidean distance.
5. The method for predicting the carbon content and the temperature of the molten steel at the end point of the converter according to claim 3, wherein the calculating the similarity of the time sequence type influence factors of the historical case and the problem case comprises: and calculating the similarity of the time sequence type influence factors of the historical cases and the problem cases based on a dynamic time warping algorithm.
6. The method for predicting carbon content and temperature of converter end point molten steel according to claim 1, wherein the step of obtaining predicted values of carbon content and temperature of converter end point molten steel corresponding to the problem case by reusing cases based on the retrieved similar cases and the carbon content and temperature of converter end point molten steel corresponding to the similar cases comprises the steps of:
and solving by adopting a k nearest neighbor method based on the retrieved similar cases and the carbon content and the temperature of the molten steel at the converter end point corresponding to the similar cases to obtain a predicted value of the carbon content and the temperature of the molten steel at the converter end point corresponding to the problem case.
7. A device for predicting carbon content and temperature of molten steel at a converter terminal point is characterized by comprising:
the case description module is used for respectively regarding each converter smelting process as a case and describing the converter smelting processes; the case description comprises case characteristic description and case solution description, and the case characteristics comprise single-value type influence factors and time sequence type influence factors which influence the carbon content and the temperature of molten steel at the converter end point; the solution comprises the carbon content and the temperature of the molten steel at the end point of the converter in the corresponding converter smelting process;
the case retrieval module is used for retrieving similar cases and converter end point molten steel carbon contents and temperatures of the similar cases, wherein the similarity between the similar cases and the problem cases in the historical cases meets preset requirements, and the converter end point molten steel carbon contents and temperatures of the similar cases are obtained based on case characteristic description of the case description module on the problem cases by taking the current converter smelting process as the problem cases and the historical converter smelting process as the historical cases;
and the case reuse module is used for obtaining the predicted values of the carbon content and the temperature of the converter end point molten steel corresponding to the problem case through case reuse based on the similar cases searched by the case search module and the converter end point molten steel corresponding to the similar cases.
8. The apparatus for predicting carbon content and temperature of molten steel at the end point of a converter according to claim 7, wherein the single value type influencing factors include molten iron information, an amount of addition of auxiliary materials, and a gas consumption amount; wherein the content of the first and second substances,
the molten iron information comprises molten iron temperature, molten iron weight and molten iron composition; the molten iron components comprise molten iron carbon content, molten iron silicon content, molten iron manganese content, molten iron phosphorus content and molten iron sulfur content;
the auxiliary raw materials comprise scrap steel, lime, light-burned dolomite and sinter;
the gas consumption comprises oxygen consumption and argon consumption;
the time sequence type influencing factors comprise oxygen supply flow, oxygen lance position and bottom argon blowing flow.
9. The apparatus of claim 7, wherein the case retrieval module is specifically configured to:
calculating the similarity of the single-value type influence factors of the historical cases and the problem cases;
calculating the similarity of time sequence type influence factors of the historical cases and the problem cases;
carrying out weighted combination on the similarity of the single-value type influence factors and the similarity of the sequence type influence factors of the historical cases and the problem cases to obtain the comprehensive similarity between the corresponding historical cases and the problem cases;
according to the comprehensive similarity, screening out similar cases meeting preset requirements from the historical cases;
the calculating the similarity of the single-value type influence factors of the historical cases and the problem cases comprises the following steps: calculating the similarity of single-value type influence factors of the historical case and the problem case based on the Euclidean distance; the calculating the similarity of the time sequence type influence factors of the historical cases and the problem cases comprises the following steps: and calculating the similarity of the time sequence type influence factors of the historical cases and the problem cases based on a dynamic time warping algorithm.
10. The apparatus of claim 7, wherein the case reuse module is specifically configured to:
and solving by adopting a k nearest neighbor method based on the retrieved similar cases and the carbon content and the temperature of the molten steel at the converter end point corresponding to the similar cases to obtain a predicted value of the carbon content and the temperature of the molten steel at the converter end point corresponding to the problem case.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110444384.6A CN113255102B (en) | 2021-04-23 | 2021-04-23 | Method and device for predicting carbon content and temperature of molten steel at converter end point |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110444384.6A CN113255102B (en) | 2021-04-23 | 2021-04-23 | Method and device for predicting carbon content and temperature of molten steel at converter end point |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113255102A true CN113255102A (en) | 2021-08-13 |
| CN113255102B CN113255102B (en) | 2022-02-08 |
Family
ID=77221452
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110444384.6A Active CN113255102B (en) | 2021-04-23 | 2021-04-23 | Method and device for predicting carbon content and temperature of molten steel at converter end point |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255102B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111104856A (en) * | 2019-11-18 | 2020-05-05 | 中冶赛迪技术研究中心有限公司 | Converter smelting splash monitoring method, system, storage medium and equipment |
| CN113656930A (en) * | 2021-06-24 | 2021-11-16 | 华北理工大学 | Method for predicting phosphorus content of smelting end point by single slag method |
| CN115456264A (en) * | 2022-09-01 | 2022-12-09 | 北京科技大学 | Method for predicting end point carbon content and end point temperature of small and medium-sized converters |
| CN115659709A (en) * | 2022-12-28 | 2023-01-31 | 湖南华联云创信息科技有限公司 | Method and system for predicting molten steel temperature in refining process of LF (ladle furnace) refining furnace |
| CN116030900A (en) * | 2023-03-24 | 2023-04-28 | 安徽瑞邦数科科技服务有限公司 | Method, device, equipment and storage medium for controlling component content of chemical product |
| CN116434856A (en) * | 2023-03-17 | 2023-07-14 | 北京科技大学 | Converter oxygen supply prediction method based on sectional oxygen decarburization efficiency |
| CN117572914A (en) * | 2023-11-17 | 2024-02-20 | 北京科技大学 | LF refined molten steel temperature control method and device based on interpretable machine learning |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102163261A (en) * | 2011-04-08 | 2011-08-24 | 汪红兵 | Case-reasoning-based molten steel temperature prediction method |
| KR20140017180A (en) * | 2012-07-31 | 2014-02-11 | 현대제철 주식회사 | Refining method for molten steel in converter |
| CN105925750A (en) * | 2016-05-13 | 2016-09-07 | 南阳理工学院 | Steelmaking end point prediction method based on neural networks |
| CN109975507A (en) * | 2019-04-28 | 2019-07-05 | 华北理工大学 | A kind of real-time determining method and system for making steel later period carbon content of molten steel and temperature value |
| CN111020118A (en) * | 2019-12-25 | 2020-04-17 | 武汉科技大学 | RH endpoint temperature prediction method based on particle swarm optimization case reasoning |
| CN111334636A (en) * | 2020-03-27 | 2020-06-26 | 宝钢湛江钢铁有限公司 | Converter end point dynamic control method and system based on sublance detection |
-
2021
- 2021-04-23 CN CN202110444384.6A patent/CN113255102B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102163261A (en) * | 2011-04-08 | 2011-08-24 | 汪红兵 | Case-reasoning-based molten steel temperature prediction method |
| KR20140017180A (en) * | 2012-07-31 | 2014-02-11 | 현대제철 주식회사 | Refining method for molten steel in converter |
| CN105925750A (en) * | 2016-05-13 | 2016-09-07 | 南阳理工学院 | Steelmaking end point prediction method based on neural networks |
| CN109975507A (en) * | 2019-04-28 | 2019-07-05 | 华北理工大学 | A kind of real-time determining method and system for making steel later period carbon content of molten steel and temperature value |
| CN111020118A (en) * | 2019-12-25 | 2020-04-17 | 武汉科技大学 | RH endpoint temperature prediction method based on particle swarm optimization case reasoning |
| CN111334636A (en) * | 2020-03-27 | 2020-06-26 | 宝钢湛江钢铁有限公司 | Converter end point dynamic control method and system based on sublance detection |
Non-Patent Citations (2)
| Title |
|---|
| HUI LIU 等: "Basic oxygen furnace steelmaking end-point prediction based on computer vision and general regression neural network", 《OPTIK》 * |
| 史战东: "转炉终点控制模型的比较分析和改进研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅰ辑》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111104856A (en) * | 2019-11-18 | 2020-05-05 | 中冶赛迪技术研究中心有限公司 | Converter smelting splash monitoring method, system, storage medium and equipment |
| CN111104856B (en) * | 2019-11-18 | 2023-06-30 | 中冶赛迪技术研究中心有限公司 | Converter smelting splash monitoring method, system, storage medium and equipment |
| CN113656930A (en) * | 2021-06-24 | 2021-11-16 | 华北理工大学 | Method for predicting phosphorus content of smelting end point by single slag method |
| CN113656930B (en) * | 2021-06-24 | 2023-11-17 | 华北理工大学 | Prediction method for smelting endpoint phosphorus content by adopting machine learning algorithm |
| CN115456264A (en) * | 2022-09-01 | 2022-12-09 | 北京科技大学 | Method for predicting end point carbon content and end point temperature of small and medium-sized converters |
| CN115659709A (en) * | 2022-12-28 | 2023-01-31 | 湖南华联云创信息科技有限公司 | Method and system for predicting molten steel temperature in refining process of LF (ladle furnace) refining furnace |
| CN116434856A (en) * | 2023-03-17 | 2023-07-14 | 北京科技大学 | Converter oxygen supply prediction method based on sectional oxygen decarburization efficiency |
| CN116434856B (en) * | 2023-03-17 | 2024-04-16 | 北京科技大学 | Converter oxygen supply prediction method based on sectional oxygen decarburization efficiency |
| CN116030900A (en) * | 2023-03-24 | 2023-04-28 | 安徽瑞邦数科科技服务有限公司 | Method, device, equipment and storage medium for controlling component content of chemical product |
| CN117572914A (en) * | 2023-11-17 | 2024-02-20 | 北京科技大学 | LF refined molten steel temperature control method and device based on interpretable machine learning |
| CN117572914B (en) * | 2023-11-17 | 2024-04-16 | 北京科技大学 | LF refined molten steel temperature control method and device based on interpretable machine learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113255102B (en) | 2022-02-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113255102B (en) | Method and device for predicting carbon content and temperature of molten steel at converter end point | |
| Cemernek et al. | Machine learning in continuous casting of steel: A state-of-the-art survey | |
| CN114611844B (en) | Method and system for determining alloy addition amount in converter tapping process | |
| Liu et al. | A stacked autoencoder with sparse Bayesian regression for end-point prediction problems in steelmaking process | |
| CN114678080B (en) | Converter end point phosphorus content prediction model, construction method and phosphorus content prediction method | |
| CN110796270A (en) | Machine learning model selection method | |
| An et al. | A multi-time-scale fusion prediction model for the gas utilization rate in a blast furnace | |
| WO2021004198A1 (en) | Plate performance prediction method and apparatus | |
| CN115034465B (en) | Coke quality prediction method based on artificial intelligence | |
| Ordieres-Meré et al. | Comparison of models created for the prediction of the mechanical properties of galvanized steel coils | |
| Wang et al. | Strip hardness prediction in continuous annealing using multiobjective sparse nonlinear ensemble learning with evolutionary feature selection | |
| Gu et al. | An improved CBR model using time-series data for predicting the end-point of a converter | |
| Wang et al. | Whale optimization end-point control model for 260 tons BOF steelmaking | |
| Tang et al. | An estimation of distribution algorithm with resampling and local improvement for an operation optimization problem in steelmaking process | |
| Li et al. | Physical metallurgy guided industrial big data analysis system with data classification and property prediction | |
| CN112418522B (en) | Industrial heating furnace steel temperature prediction method based on three-branch integrated prediction model | |
| Liu et al. | End-point prediction of 260 tons basic oxygen furnace (BOF) steelmaking based on WNPSVR and WOA | |
| CN115456264B (en) | Endpoint carbon content and endpoint temperature prediction method for small and medium-sized converter | |
| CN117093868A (en) | Converter endpoint prediction method and device based on multitask learning | |
| Zhang et al. | Deep learning-based prediction framework of temperature control time for wide-thick slab hot rolling production | |
| Sala et al. | Hybrid static-sensory data modeling for prediction tasks in basic oxygen furnace process | |
| Han et al. | BOF oxygen control by mixed case retrieve and reuse CBR | |
| CN116484745A (en) | Design method of hot-rolled low-alloy steel for physical metallurgy guidance industry big data mining | |
| Hu et al. | Blast Furnace Thermal State Prediction Based on Multiobjective Evolutionary Ensemble Neural Networks | |
| Shi et al. | Prediction Model of Yield Strength of V–N Steel Hot-rolled Plate Based on Machine Learning Algorithm |
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 |