An intelligent adjustment method for blast furnace slag basicity based on big data

Abstract

The invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which belongs to the technical field of intelligent manufacturing. The method comprises the following steps: collecting historical production data of the blast furnace, determining key production parameters according to the historical production data, and generating a theoretical calculation model for the basicity of the blast furnace slag; determining target parameters related to the basicity of the slag in combination with process principles, screening the relevant target parameters, obtaining strongly correlated parameters, and constructing a prediction model; performing a state evaluation on the blast furnace slag iron based on the prediction result, obtaining a state evaluation result, and determining an adjustment direction and an adjustment step length of the basicity of the slag according to the state evaluation result; determining a material adjustment amount based on the theoretical calculation model and current furnace charge data, combining the adjustment direction and the adjustment step length of the basicity of the slag, and intelligently adjusting the basicity of the slag one smelting cycle in advance, so as to achieve efficient and accurate control of the basicity of the slag, reduce manual intervention, optimize the operation process, keep the basicity within an optimal range, and reduce the unqualified rate.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing technology, and in particular to a method for intelligently adjusting blast furnace slag basicity based on big data. Background Art

[0002] Slag basicity control is an important part of the ironmaking process. The basicity control process has the characteristics of long time lag, many influencing factors, and strong parameter coupling. In the actual production process, on-site operators usually observe the slag parameter data based on the existing equipment conditions and judge whether there are abnormalities in the smelting process state parameters. They adjust the blast furnace raw fuel parameters through theoretical calculations combined with their own experience. The reaction speed is slow, the labor intensity is high, and the recovery period of slag quality abnormalities is long, resulting in low efficiency of the ironmaking process.

[0003] Therefore, the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data. Summary of the invention

[0004] The present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data. The method collects historical production data of the blast furnace, establishes a theoretical calculation model for the basicity of the slag, screens out strongly correlated parameters in combination with process principles, constructs a prediction model, performs status evaluation based on the prediction results, determines the adjustment direction and step size, and finally dynamically adjusts the material through an intelligent algorithm, thereby achieving efficient and accurate control of the basicity of the slag, reducing manual intervention, optimizing the operating process, keeping the basicity within the optimal range, and reducing the unqualified rate.

[0005] The present invention provides a method for intelligently adjusting blast furnace slag basicity based on big data, comprising:

[0006] Step 1: Collect historical production data of blast furnaces, determine key production parameters based on the historical production data, and generate a theoretical calculation model for blast furnace slag basicity;

[0007] Step 2: Determine the target parameters related to slag basicity in combination with the process principles, screen the related target parameters, obtain strong correlation parameters, and build a prediction model based on the strong correlation parameters;

[0008] Step 3: Based on the prediction results, the actual data is evaluated to obtain the evaluation results, and the adjustment direction and step length of the actual slag basicity are determined according to the evaluation results;

[0009] Step 4: Based on the theoretical calculation model and the current charge data, the material adjustment amount is determined in combination with the actual adjustment direction and actual adjustment step of the actual slag basicity, and the current actual slag basicity is intelligently adjusted;

[0010] In step 2, the actual situation of the ironmaking process is obtained, and the process principle is determined based on the actual situation, thereby identifying the first relevant target parameter of slag basicity and the second relevant target parameter fed back by the operator, and calculating the correlation between these two parameters and the slag basicity, constructing a correlation matrix and performing preliminary screening, performing collinearity analysis to merge the parameters, and finally obtaining strongly correlated parameters, and constructing a prediction model based on the strongly correlated parameters;

[0011] In step 3, the parameter ranges of slag basicity, magnesium-aluminum ratio and molten iron sulfur content are set according to field experience and product quality standards, and state classification is performed accordingly. After obtaining the actual data, the actual state classification corresponding to the actual data is determined, and the actual state classification is compared with the state classification of the predicted results to derive the adjustment direction of the actual slag basicity and calculate the corresponding adjustment step.

[0012] The present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, collects historical production data of the blast furnace, determines key production parameters according to the historical production data, and generates a theoretical calculation model for the basicity of blast furnace slag, including:

[0013] Collecting historical production data of a blast furnace, drawing a first box plot using statistical software according to the historical production data, and processing abnormal values ​​in the historical production data according to the first box plot;

[0014] Identify key production parameters in the processing result according to the first box plot, and draw a second box plot according to the key production parameters to identify parameter distribution characteristics;

[0015] Draw a heat map for the processing results to determine the connection between each key production parameter and the slag basicity, and determine the influence of different raw material ratios on the slag basicity based on the parameter distribution characteristics and the connection relationship;

[0016] An initial model is constructed according to the degree of influence combined with a preset basicity adjustment principle. According to the on-site material intake, the current values ​​of key production parameters are input into the initial model to generate a theoretical calculation model for blast furnace slag basicity.

[0017] The present invention provides a method for intelligently adjusting blast furnace slag basicity based on big data, which determines target parameters related to slag basicity in combination with process principles, screens the relevant target parameters, obtains strongly correlated parameters, and constructs a prediction model based on the strongly correlated parameters, including:

[0018] Acquire the actual situation of the ironmaking process, determine the process principle based on the actual situation, determine the first relevant target parameter of slag basicity based on the process principle, collect operator feedback, and obtain the second relevant target parameter of slag basicity;

[0019] Calculating the correlation between the first related target parameter, the second related target parameter and the slag basicity, constructing a correlation matrix, performing a first screening on the first related target parameter and the second related target according to the correlation matrix, and obtaining a first screening result;

[0020] Performing a collinearity analysis on the first relevant target parameter and the second relevant target parameter, and merging the first relevant target parameter and the second relevant target parameter according to the collinearity analysis result to obtain a second screening result;

[0021] The first screening result and the second screening result are combined to obtain strong correlation parameters, and then a prediction model is constructed.

[0022] The present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which combines the first screening result and the second screening result to obtain a strong correlation parameter, and then constructs a prediction model, including:

[0023] Taking blast furnace slag basicity adjustment as the target to be solved, the output type of the prediction model is determined, and the input features are selected from the strongly associated parameters according to the target-feature table;

[0024] Based on the input features, a deep learning algorithm is used to construct a prediction model for predicting the parameters slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content one smelting cycle in advance.

[0025] The present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which performs state evaluation on actual data based on prediction results, obtains state evaluation results, and determines the adjustment direction and adjustment step length of the actual slag basicity according to the state evaluation results, including:

[0026] According to field experience and product quality standards, parameter ranges corresponding to the parameter slag basicity, parameter slag magnesium-aluminum ratio and parameter molten iron sulfur content are set, and the parameter ranges are used to perform state assessment, and then the state classification is determined according to the state assessment results;

[0027] Acquire actual data, determine the actual state classification corresponding to the actual data according to the state classification, compare the actual state classification with the predicted state classification corresponding to the predicted result, derive the adjustment direction of the actual slag basicity, and calculate the adjustment step size corresponding to the adjustment direction.

[0028] The present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which calculates the adjustment step length corresponding to the adjustment direction, including:

[0029] Calculate the adjustment step size:

[0030] ,in, Indicates the adjustment step of actual slag basicity; K indicates the basic adjustment coefficient; Indicates the actual slag basicity and prediction of slag basicity The influence function of the gap between them on the step length; represents the historical step-size feedback function of the parameter slag basicity; represents the historical adjustment feedback coefficient; Indicates the fluctuation of the parameters slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content in the current time step t; It represents the average value of the parameter slag basicity, parameter slag magnesium-aluminum ratio and parameter molten iron sulfur content in the current time step t; Represents the adjustment coefficient of the coupling factor to the step size; It represents the coupling factor function of the parameter slag basicity, the parameter slag magnesium-aluminum ratio and the parameter molten iron sulfur content; t represents the current time step; T represents the total adjustment period; S represents the stability constant of the smelting process; i represents the parameter slag basicity; j represents the parameter slag magnesium-aluminum ratio; m represents the parameter molten iron sulfur content.

[0031] The present invention provides a method for intelligently adjusting blast furnace slag basicity based on big data, a coupling factor function, comprising:

[0032] ,in, It represents the coupling factor function between the parameter slag basicity i and the parameter slag magnesium-aluminum ratio j; represents the actual slag basicity at time step k; Indicates the actual slag basicity The mean value in time step t; It represents the actual slag magnesium-aluminum ratio at time step k; Indicates the actual slag magnesium-aluminum ratio The mean value in time step t; k0 represents the time window size; k represents the time step;

[0033] ,in, It represents the coupling factor function between the parameter slag basicity i and the parameter molten iron sulfur content m; represents the actual sulfur content of molten iron at time step k; Indicates the actual sulfur content of molten iron The mean value in time step t;

[0034] ,in, It represents the coupling factor function between the parameter molten iron sulfur content m and the parameter slag magnesium-aluminum ratio j;

[0035] ,in, The weight coefficient representing the coupling relationship between the parameter slag basicity i and the parameter slag magnesium-aluminum ratio j; The weight coefficient representing the coupling relationship between the parameter slag basicity i and the parameter molten iron sulfur content m; It represents the weight coefficient of the coupling relationship between the parameter molten iron sulfur content m and the parameter slag magnesium-aluminum ratio j.

[0036] The present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data. Based on a theoretical calculation model and current charging data, the material adjustment amount is determined in combination with the actual adjustment direction and the actual adjustment step of the actual slag basicity, and the current actual slag basicity is intelligently adjusted, including:

[0037] Obtaining current furnace charge data, inputting the current furnace charge data into a theoretical calculation model, and calculating the current ideal slag basicity under ideal conditions under the current furnace charge data according to the theoretical calculation model;

[0038] Input the current charge data into the prediction model, calculate the current predicted slag basicity under the current charge data according to the prediction model, and then obtain the actual adjustment direction and actual adjustment step of the current actual slag basicity;

[0039] The material adjustment amount is determined based on the adjustment direction and adjustment step of the current ideal slag basicity and the current actual slag basicity, an adjustment instruction is generated according to the material adjustment amount, and the generated adjustment instruction is passed to the blast furnace operating system to make corresponding material adjustments, thereby intelligently adjusting the current actual slag basicity.

[0040] Compared with the prior art, the beneficial effects of the present application are as follows: by collecting historical production data of blast furnaces, a theoretical calculation model of slag basicity is established, and strongly correlated parameters are screened out in combination with process principles, a prediction model is constructed, and status evaluation is performed based on the prediction results to determine the adjustment direction and step size. Finally, the material is dynamically adjusted through an intelligent algorithm, thereby achieving efficient and accurate slag basicity control, reducing manual intervention, optimizing operating procedures, keeping the basicity within the optimal range, and reducing the unqualified rate.

[0041] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0042] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0044] Figure 1 It is a flow chart of a method for intelligently adjusting the basicity of blast furnace slag based on big data provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0046] Embodiment 1:

[0047] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, such as Figure 1 As shown, including:

[0048] Step 1: Collect historical production data of blast furnaces, determine key production parameters based on the historical production data, and generate a theoretical calculation model for blast furnace slag basicity;

[0049] Step 2: Determine the target parameters related to slag basicity in combination with the process principles, screen the related target parameters, obtain strong correlation parameters, and build a prediction model based on the strong correlation parameters;

[0050] Step 3: Based on the prediction results, the actual data is evaluated to obtain the evaluation results, and the adjustment direction and step length of the actual slag basicity are determined according to the evaluation results;

[0051] Step 4: Based on the theoretical calculation model and the current charge data, the material adjustment amount is determined in combination with the actual adjustment direction and actual adjustment step of the actual slag basicity, and the current actual slag basicity is intelligently adjusted;

[0052] In step 2, the actual situation of the ironmaking process is obtained, and the process principle is determined based on the actual situation, thereby identifying the first relevant target parameter of slag basicity and the second relevant target parameter fed back by the operator, and calculating the correlation between these two parameters and the slag basicity, constructing a correlation matrix and performing preliminary screening, performing collinearity analysis to merge the parameters, and finally obtaining strongly correlated parameters, and constructing a prediction model based on the strongly correlated parameters;

[0053] In step 3, the parameter ranges of slag basicity, magnesium-aluminum ratio and molten iron sulfur content are set according to field experience and product quality standards, and state classification is performed accordingly. After obtaining the actual data, the actual state classification corresponding to the actual data is determined, and the actual state classification is compared with the state classification of the predicted results to derive the adjustment direction of the actual slag basicity and calculate the corresponding adjustment step.

[0054] In this embodiment, historical production data refers to the production records of the blast furnace in the past period of time, including temperature, pressure, raw material ratio, slag basicity and other related parameters, for example, the temperature, feed rate, slag basicity and other data of each blast furnace operation in the past year.

[0055] In this embodiment, the key production parameters are important indicators that have a great impact on the blast furnace production process, such as temperature, raw material ratio, slag basicity, gas flow rate, etc.

[0056] In this embodiment, the theoretical calculation model is a model constructed by material balance using blast furnace raw fuel data and slag iron data. Function 1 is to calculate the theoretical basicity of slag in real time, and function 2 is used for feedback calculation of material adjustment amount.

[0057] In this embodiment, by collecting historical production data of the blast furnace, using statistical software to draw box plots, identifying and processing outliers, analyzing the distribution characteristics of key parameters, and drawing heat maps to determine the relationship between these parameters and slag basicity, based on these analysis results and combined with the preset basicity adjustment principles, an initial model is constructed, and the current key production parameters are input into the model to generate a theoretical calculation model for blast furnace slag basicity.

[0058] In this embodiment, the process principles are the basic rules and operating specifications followed in the ironmaking process to ensure production efficiency and product quality, for example, ensuring that the slag basicity is within an appropriate range to optimize metal recovery and reduce energy consumption.

[0059] In this embodiment, the relevant target parameters include a first relevant target parameter and a second relevant target parameter. The first relevant target parameter is a key operating parameter directly related to the slag basicity, which is usually determined according to process principles, such as furnace temperature and raw material ratio (such as the ratio of coke, iron ore, and limestone). The second relevant target parameter is other key parameters determined based on operator feedback, which may indirectly affect the slag basicity, such as gas flow rate, furnace pressure, reaction time, etc.

[0060] In this embodiment, the screening process includes a first screening and a second screening. The first screening is based on correlation evaluation to screen out parameters that are significantly correlated with slag basicity. The second screening is based on collinearity analysis to further merge and simplify parameters to ensure the effectiveness and stability of the model. For example, assuming that in the first screening result, the VIF value between "charge composition (Fe content)" and "charge composition (CaO content)" is 12, indicating that there is a strong collinearity between them. According to the collinearity analysis results, "charge composition (Fe content)" can be selected as the representative parameter, and "charge composition (CaO content)" can be merged to obtain the second screening result.

[0061] In this embodiment, the strongly correlated parameters are the parameters that are determined to have the strongest relationship with the slag basicity after the first screening and the second screening, for example, the furnace temperature and the raw material ratio are finally determined.

[0062] In this embodiment, the prediction model input includes strongly correlated parameters, for example, the input parameters are furnace temperature and raw material ratio, and the output is the prediction parameters of slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content one smelting cycle in advance, and the number of training times is 10,000.

[0063] In this embodiment, the output type of the prediction model is determined by adjusting the basicity of blast furnace slag as the goal, and input features are selected from strongly correlated parameters through the target-feature table. A model is constructed using a deep learning algorithm to predict slag basicity, magnesium-aluminum ratio and molten iron sulfur content one smelting cycle in advance, analyze historical data, capture the complex relationship between parameters, and generate accurate prediction results.

[0064] In this embodiment, the status assessment result refers to the assessment result on blast furnace slag iron obtained after comprehensive analysis of key parameters such as slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content.

[0065] In this embodiment, the adjustment direction is to determine the adjustment measures that need to be taken based on the comparison results to optimize the production process. For example, if the actual slag basicity is 1.17, but the target is 1.2, the adjustment direction is to "increase" the slag basicity. If the actual molten iron sulfur content is 0.72%, the adjustment direction is to "reduce" the molten iron sulfur content.

[0066] In this embodiment, the adjustment step size is the amplitude of each adjustment set according to the adjustment direction, which is usually based on experience or historical data. For example, if it is decided to increase the slag basicity, the adjustment step size is set to 0.01, which means that the target after each adjustment is 1.18.

[0067] In this embodiment, the current charge data refers to the raw material data currently put into the blast furnace, including the composition, quantity and properties of ore, coke, additives, etc. For example, the current ore data entering the furnace may include: iron ore (Fe content 30%), coke (C content 10%), limestone (CaO content 5%), etc.

[0068] In this embodiment, the material adjustment amount is the amount of raw material that needs to be increased or decreased according to the actual adjustment direction and the actual adjustment step. For example, if the slag basicity needs to be increased by 0.03, a certain amount of limestone may need to be added (for example, 100 kg of limestone).

[0069] In this embodiment, intelligent adjustment is achieved by predicting slag basicity, molten iron sulfur and slag magnesium-aluminum ratio through a prediction model one smelting cycle in advance, determining the adjustment direction and step size of the slag basicity, and then obtaining the current furnace charge data and inputting it into the theoretical calculation model to calculate the material adjustment amount. The generated adjustment instruction is passed to the blast furnace operating system to realize intelligent adjustment of the actual slag basicity.

[0070] The working principle and beneficial effects of the above technical solution are: by collecting historical production data of blast furnaces, establishing a theoretical calculation model for slag basicity, and combining process principles to screen out strongly correlated parameters, constructing a prediction model, and conducting status assessment based on the prediction results, determining the adjustment direction and step size, and finally dynamically adjusting the material through an intelligent algorithm, thereby achieving efficient and accurate slag basicity control, reducing manual intervention, optimizing operating procedures, keeping the basicity within the optimal range, and reducing the unqualified rate.

[0071] Embodiment 2:

[0072] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, collects historical production data of the blast furnace, determines key production parameters according to the historical production data, and generates a theoretical calculation model for the basicity of blast furnace slag, including:

[0073] Collecting historical production data of a blast furnace, drawing a first box plot using statistical software according to the historical production data, and processing abnormal values ​​in the historical production data according to the first box plot;

[0074] Identify key production parameters in the processing result according to the first box plot, and draw a second box plot according to the key production parameters to identify parameter distribution characteristics;

[0075] Draw a heat map for the processing results to determine the connection between each key production parameter and the slag basicity, and determine the influence of different raw material ratios on the slag basicity based on the parameter distribution characteristics and the connection relationship;

[0076] An initial model is constructed according to the degree of influence combined with a preset basicity adjustment principle. According to the on-site material intake, the current values ​​of key production parameters are input into the initial model to generate a theoretical calculation model for blast furnace slag basicity.

[0077] In this embodiment, the statistical software is a tool for data analysis, drawing and statistical calculation, such as R, Python (using Pandas and Matplotlib libraries), SPSS or Minitab, etc.

[0078] In this embodiment, the first box plot is a visualization tool for displaying the distribution of data, including the median, quartiles and outliers. For example, a box plot of blast furnace slag basicity is drawn to display its distribution range and outliers.

[0079] In this embodiment, outlier processing is to identify and process values ​​that deviate significantly from the normal range in data analysis. For example, in the first box plot, if a slag basicity value exceeds 1.5 times the interquartile range (IQR), it is regarded as an outlier and can be deleted or replaced with the median.

[0080] In this embodiment, the processing result is a cleaned data set obtained after outlier processing, for example, a cleaned data set of blast furnace slag basicity is obtained after removing outliers.

[0081] In this embodiment, the second box plot is drawn based on the processed key production parameter to identify its distribution characteristics. For example, a box plot of the processed temperature is drawn to observe its distribution and abnormal values.

[0082] In this embodiment, the heat map is a visualization tool for displaying the correlation between multiple variables, and for showing the connection and influence between different key production parameters (such as furnace temperature, raw material ratio, gas flow rate, etc.) and slag basicity. For example, a heat map is drawn to show the correlation between slag basicity and parameters such as furnace temperature, coke ratio, and iron ore ratio. The color of the heat map can indicate the size of the correlation coefficient. Dark colors may indicate strong correlation, and light colors indicate weak correlation.

[0083] In this embodiment, the initial model is a theoretical model established based on parameter distribution characteristics and connection relationships. For example, a linear regression model is used to establish a calculation model for slag basicity according to the treatment results and basicity adjustment principles.

[0084] In this embodiment, the parameter distribution characteristics refer to the distribution of a certain parameter within a certain range, including mean, standard deviation, skewness, peak state, etc. For example, the distribution characteristics of the blast furnace temperature may show a mean of 1500°C, a standard deviation of 50°C, and a normal distribution.

[0085] In this embodiment, the connection relationship refers to the correlation or causal relationship between different parameters. For example, there may be a positive correlation between slag basicity and furnace temperature, that is, the higher the temperature, the higher the slag basicity tends to be.

[0086] In this embodiment, the degree of influence refers to the influence of a certain factor on the target variable, which is usually quantified through statistical analysis. For example, the change in raw material ratio may affect the basicity of the slag to the extent that for every 1% increase in alkaline raw materials, the basicity of the slag increases by 0.05 units.

[0087] In this embodiment, the preset basicity adjustment principle refers to the guiding principle for adjusting the slag basicity during the production process. For example, if the slag basicity is lower than 1.19, the proportion of alkaline raw materials is increased; if it is higher than 1.23, the proportion of alkaline raw materials is reduced.

[0088] In this embodiment, the input of the initial model is the parameters used for prediction, and the output is the prediction result. The input includes key production parameters such as current furnace temperature, raw material ratio, gas flow rate, theoretically calculated slag basicity, and the number of training times is 10,000.

[0089] In this embodiment, the on-site material consumption refers to the usage and proportion of raw materials in the actual production process of the blast furnace. For example, it is recorded that in a certain production cycle, a proportion of 60% coke, 30% iron ore, and 10% limestone was used.

[0090] In this embodiment, the theoretical calculation model is based on physical, chemical reaction equations and thermodynamic principles, which usually requires more professional knowledge and theoretical support and may involve complex mathematical formulas and calculations; the initial model is based on statistical analysis of historical data, commonly using regression analysis, machine learning and other methods, the construction process is relatively simple, and depends on the availability and quality of data; the theoretical calculation model does not rely on historical data, but relies more on theoretical foundations and experimental data to verify the accuracy of the model; the initial model is strongly dependent on historical production data, and discovers patterns and trends through data mining and analysis.

[0091] The working principle and beneficial effects of the above technical solution are: by collecting historical production data of blast furnaces, using statistical software to draw box plots, identifying and processing outliers, analyzing the distribution characteristics of key parameters, and drawing heat maps to determine the relationship between these parameters and slag basicity, based on these analysis results, combined with preset basicity adjustment principles, an initial model is constructed, and the current key production parameters are input into the model to generate a theoretical calculation model for blast furnace slag basicity, clarify the main factors affecting slag basicity, reduce manual intervention, respond quickly to changes, and improve overall efficiency and product quality.

[0092] Embodiment 3:

[0093] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which determines the target parameters related to the basicity of slag in combination with the process principles, screens the related target parameters, obtains the strongly correlated parameters, and constructs a prediction model based on the strongly correlated parameters, including:

[0094] Acquire the actual situation of the ironmaking process, determine the process principle based on the actual situation, determine the first relevant target parameter of slag basicity based on the process principle, collect operator feedback, and obtain the second relevant target parameter of slag basicity;

[0095] Calculating the correlation between the first related target parameter, the second related target parameter and the slag basicity, constructing a correlation matrix, performing a first screening on the first related target parameter and the second related target according to the correlation matrix, and obtaining a first screening result;

[0096] Performing a collinearity analysis on the first relevant target parameter and the second relevant target parameter, and merging the first relevant target parameter and the second relevant target parameter according to the collinearity analysis result to obtain a second screening result;

[0097] The first screening result and the second screening result are combined to obtain strong correlation parameters, and then a prediction model is constructed.

[0098] In this embodiment, the actual situation of the ironmaking process refers to the actual data and conditions such as operating conditions, equipment performance, raw material characteristics, etc. observed in the actual ironmaking process, such as the feed ratio of the blast furnace, furnace temperature, gas flow rate, slag composition, etc.

[0099] In this embodiment, the operator feedback is feedback from the operator's experience and observation, which is usually used to verify and supplement the theoretical model. For example, the operator may provide feedback on the change of slag basicity at different furnace temperatures.

[0100] In this embodiment, correlation describes the strength and direction of the relationship between one parameter and another parameter. For example, the correlation between furnace temperature and slag basicity may be a positive correlation, that is, when the furnace temperature increases, the slag basicity also increases.

[0101] In this embodiment, the correlation matrix is ​​a table that displays the correlation coefficients between multiple parameters, and the Pearson correlation coefficient is usually used. For example, the first row and the first column are furnace temperature, raw material ratio, gas flow rate, and slag basicity. The second row and the second column are 1, indicating that the Pearson correlation coefficient between the furnace temperature and the furnace temperature is 1. The second row and the third column are 0.8, indicating that the Pearson correlation coefficient between the furnace temperature and the raw material ratio is 0.8.

[0102] In this embodiment, the first screening result is the parameter with a strong relationship with the slag basicity screened out based on the correlation matrix. For example, if the furnace temperature and the raw material ratio have a high correlation with the slag basicity, these two parameters will be screened out.

[0103] In this embodiment, collinearity analysis is to analyze the correlation between multiple independent variables to determine whether there is a multicollinearity problem. For example, if the correlation between furnace temperature and raw material ratio is very high, it may cause collinearity problems.

[0104] In this embodiment, the collinearity analysis result is the output of the collinearity analysis, which usually includes indicators such as variance inflation factor (VIF) and is used to determine the degree of collinearity between variables. For example, a VIF value greater than 10 may indicate severe collinearity.

[0105] In this embodiment, the second screening result is the result of merging or screening the first relevant target parameter and the second relevant target parameter after collinearity analysis. For example, if there is collinearity between furnace temperature and raw material ratio, only furnace temperature may be retained as a key parameter.

[0106] In this embodiment, the merging is to merge highly correlated parameters to reduce redundancy and collinearity effects, for example, furnace temperature and raw material ratio are merged into a comprehensive indicator.

[0107] The working principle and beneficial effects of the above technical solution are: by obtaining the actual conditions of the ironmaking process and determining the process principles based on these conditions, the first relevant target parameter of slag basicity and the second relevant target parameter of operator feedback are identified, and the correlation between these two parameters and the slag basicity is calculated, a correlation matrix is ​​constructed and preliminary screening is performed, and collinearity analysis is performed to merge the parameters, and finally strongly correlated parameters are obtained. A prediction model is constructed based on these parameters to accurately identify the key parameters affecting the slag basicity, shorten the abnormal recovery period of slag iron quality, and improve the overall production quality and efficiency.

[0108] Embodiment 4:

[0109] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which combines the first screening result and the second screening result to obtain a strong correlation parameter, and then constructs a prediction model, including:

[0110] Taking blast furnace slag basicity adjustment as the target to be solved, the output type of the prediction model is determined, and the input features are selected from the strongly associated parameters according to the target-feature table;

[0111] Based on the input characteristics, a prediction model is constructed to predict the parameter slag basicity, parameter slag magnesium-aluminum ratio and parameter molten iron sulfur content one smelting cycle in advance.

[0112] In this embodiment, the target to be solved is a key indicator that needs to be optimized or adjusted in a specific production process, which refers to the adjustment of blast furnace slag basicity.

[0113] In this embodiment, the output result type of the prediction model is usually one or more numerical values. In this case, the output types are slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content.

[0114] In this embodiment, the target-feature table lists the relationship between the target to be solved and the input features, and is usually used to select appropriate input features. For example, the target is the slag magnesium-aluminum ratio, and the corresponding input features are furnace temperature and limestone content.

[0115] In this embodiment, input features are used to train variables or parameters of the prediction model. These features should have a significant correlation with the target to be solved. For example, according to the target-feature table, the selected input features may include: furnace temperature, raw material ratio, gas flow rate and limestone content.

[0116] In this embodiment, the deep learning algorithm is a machine learning method that uses a multi-layer neural network for feature learning and pattern recognition. It is suitable for processing complex nonlinear relationships and large-scale data. For example, the neural network: builds a multi-layer perceptron (MLP) model, with input features such as furnace temperature, raw material ratio, etc., and outputs such as slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content.

[0117] The working principle and beneficial effects of the above technical solution are: by taking the adjustment of blast furnace slag basicity as the goal, determining the output type of the prediction model, selecting input features from strongly correlated parameters through the target-feature table, and using a deep learning algorithm to build a model to predict slag basicity, magnesium-aluminum ratio and molten iron sulfur content one smelting cycle in advance, analyzing historical data, capturing the complex relationship between parameters, generating accurate prediction results, achieving early warning, reducing waste of raw materials and energy, reducing quality instability due to parameter fluctuations, and improving product consistency.

[0118] Embodiment 5:

[0119] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which performs a state evaluation on actual data based on a prediction result, obtains a state evaluation result, and determines an adjustment direction and an adjustment step length of the actual slag basicity according to the state evaluation result, including:

[0120] According to field experience and product quality standards, parameter ranges corresponding to the parameter slag basicity, parameter slag magnesium-aluminum ratio and parameter molten iron sulfur content are set, and the parameter ranges are used to perform state assessment, and then the state classification is determined according to the state assessment results;

[0121] Acquire actual data, determine the actual state classification corresponding to the actual data according to the state classification, compare the actual state classification with the predicted state classification corresponding to the predicted result, derive the adjustment direction of the actual slag basicity, and calculate the adjustment step size corresponding to the adjustment direction.

[0122] In this embodiment, field experience and product quality standards are product quality requirements and production operation specifications set based on the operator's experience and industry standards, such as slag basicity: the ideal range is 1.17-1.27, slag magnesium-aluminum ratio: the ideal range is 2.0-3.0, and molten iron sulfur content: the ideal range is 0.6%-0.7%.

[0123] In this embodiment, the parameter range is an acceptable range set for each key parameter to ensure product quality and production efficiency, for example, slag basicity: 1.17-1.27, slag magnesium-aluminum ratio: 2.0-3.0, molten iron sulfur content: 0.6%-0.7%.

[0124] In this embodiment, the status classification is to divide the actual status into different levels according to the set parameter range for easy monitoring and adjustment. For example, the status classification: excellent: within the ideal range, good: close to the ideal range (such as slag basicity is 1.17-1.19 or 1.25-1.27), general: slightly below or slightly above the ideal range (such as slag basicity is 1.15-1.17 or 1.27-1.3), poor: far below or far above the ideal range (such as slag basicity is lower than 1.2 or higher than 1.35).

[0125] In this embodiment, the actual data is real-time data collected from the blast furnace production process, reflecting the current production status. For example, the actual measured slag basicity is 1.18, the magnesium-aluminum ratio is 2.5, and the molten iron sulfur content is 0.61%.

[0126] In this embodiment, the comparison process is to compare the actual data with the set parameter range and state classification to determine the current state. For example, the actual slag basicity of 1.18 falls within the "excellent" state classification, the actual magnesium-aluminum ratio of 2.5 is also within the "excellent" state classification, and the actual molten iron sulfur content of 0.61% is within the "excellent" state classification.

[0127] The working principle and beneficial effects of the above technical solution are: setting the parameter ranges of slag basicity, magnesium-aluminum ratio and molten iron sulfur content according to field experience and product quality standards, and performing state classification accordingly, and after obtaining actual data, determining the corresponding actual state classification, comparing the actual state classification with the state classification of the predicted results, and obtaining the adjustment direction of the actual slag basicity, and calculating the corresponding adjustment step to guide subsequent adjustment operations, quickly respond to actual state changes, and improve production flexibility.

[0128] Embodiment 6:

[0129] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which calculates the adjustment step length of the corresponding adjustment direction, including:

[0130] Calculate the adjustment step size:

[0131] ,in, Indicates the adjustment step of actual slag basicity; K indicates the basic adjustment coefficient; Indicates the actual slag basicity and prediction of slag basicity The influence function of the gap between them on the step length; represents the historical step-size feedback function of the parameter slag basicity; represents the historical adjustment feedback coefficient; Indicates the fluctuation of the parameters slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content in the current time step t; It represents the average value of the parameter slag basicity, parameter slag magnesium-aluminum ratio and parameter molten iron sulfur content in the current time step t; Represents the adjustment coefficient of the coupling factor to the step size; It represents the coupling factor function of the parameter slag basicity, the parameter slag magnesium-aluminum ratio and the parameter molten iron sulfur content; t represents the current time step; T represents the total adjustment period; S represents the stability constant of the smelting process; i represents the parameter slag basicity; j represents the parameter slag magnesium-aluminum ratio; m represents the parameter molten iron sulfur content.

[0132] In this embodiment, the influence function corresponding formula is ,in, Represents the scaling factor for adjusting the step sensitivity; A sensitivity factor that represents the impact of the previous step adjustment.

[0133] The working principle and beneficial effects of the above technical solution are: through multiple factors, the gap between the actual slag basicity and the predicted value, historical feedback, volatility and coupling factors are comprehensively considered. The basic adjustment coefficient K and the influence function determine the initial value of the step length, while the historical feedback and volatility adjust the sensitivity of the step length to ensure adaptability under different conditions. Through the dynamic coupling of these parameters, the optimal adjustment step length is calculated to achieve accurate slag basicity adjustment, reduce the number of adjustments, reduce resource consumption, and improve production efficiency.

[0134] Embodiment 7:

[0135] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, and a coupling factor function, including:

[0136] ,in, It represents the coupling factor function between the parameter slag basicity i and the parameter slag magnesium-aluminum ratio j; represents the actual slag basicity at time step k; Indicates the actual slag basicity The mean value in time step t; It represents the actual slag magnesium-aluminum ratio at time step k; Indicates the actual slag magnesium-aluminum ratio The mean value in time step t; k0 represents the time window size; k represents the time step;

[0137] ,in, It represents the coupling factor function between the parameter slag basicity i and the parameter molten iron sulfur content m; represents the actual sulfur content of molten iron at time step k; Indicates the actual sulfur content of molten iron The mean value in time step t;

[0138] ,in, It represents the coupling factor function between the parameter molten iron sulfur content m and the parameter slag magnesium-aluminum ratio j;

[0139] ,in, The weight coefficient representing the coupling relationship between the parameter slag basicity i and the parameter slag magnesium-aluminum ratio j; The weight coefficient representing the coupling relationship between the parameter slag basicity i and the parameter molten iron sulfur content m; It represents the weight coefficient of the coupling relationship between the parameter molten iron sulfur content m and the parameter slag magnesium-aluminum ratio j.

[0140] The working principle and beneficial effects of the above technical solution are: the mutual influence between the parameters of slag basicity, magnesium-aluminum ratio and molten iron sulfur content is described by a coupling factor function, each coupling factor function is calculated according to the actual value and mean of the time step k, reflecting the dynamic relationship between different parameters, and the weight coefficient is used to adjust the importance of each coupling relationship, so as to comprehensively consider the influence of each parameter on the adjustment process, so as to achieve the optimal adjustment of slag basicity, ensure that each parameter is within the optimal range, and improve the consistency and quality of the final product.

[0141] Embodiment 8:

[0142] The embodiment of the present invention provides a method for intelligently adjusting the basicity of blast furnace slag based on big data, which determines the material adjustment amount based on the theoretical calculation model and the current charge data, combined with the actual adjustment direction and the actual adjustment step of the actual slag basicity, and intelligently adjusts the current actual slag basicity, including:

[0143] Obtaining current furnace charge data, inputting the current furnace charge data into a theoretical calculation model, and calculating the current ideal slag basicity under ideal conditions under the current furnace charge data according to the theoretical calculation model;

[0144] Input the current charge data into the prediction model, calculate the current predicted slag basicity under the current charge data according to the prediction model, and then obtain the actual adjustment direction and actual adjustment step of the current actual slag basicity;

[0145] The material adjustment amount is determined based on the adjustment direction and adjustment step of the current ideal slag basicity and the current actual slag basicity, an adjustment instruction is generated according to the material adjustment amount, and the generated adjustment instruction is passed to the blast furnace operating system to make corresponding material adjustments, thereby intelligently adjusting the current actual slag basicity.

[0146] In this embodiment, the current ideal slag basicity is the ideal slag basicity value obtained by the theoretical calculation model under specific furnace charge conditions and without any faults, which represents the chemical properties of the slag under optimal conditions. For example, based on the current furnace charge data, the theoretical calculation model may conclude that the ideal slag basicity is 1.2.

[0147] In this embodiment, the current predicted slag basicity is a slag basicity value calculated using a prediction model based on current furnace charge data, taking into account historical data and the prediction capability of the model. For example, the prediction model may show that under current furnace charge conditions, the predicted slag basicity is 1.18.

[0148] In this embodiment, the current actual slag basicity is the actual basicity value of the current slag obtained through real-time monitoring, which reflects the actual operating state of the blast furnace. For example, the actual measured slag basicity is 1.16.

[0149] In this embodiment, the actual adjustment direction is determined by the difference between the ideal slag basicity and the actual slag basicity to determine the adjustment direction that needs to be increased or decreased. For example, if the ideal slag basicity is 1.2 and the actual slag basicity is 0.9, the adjustment direction is "increase".

[0150] In this embodiment, the actual adjustment step size is the step size calculated based on the deviation function, which indicates the size of the amount to be adjusted. For example, if the calculated step size is 0.02, it means that the slag basicity needs to be increased by 0.02 during the adjustment process.

[0151] In this embodiment, the adjustment instruction is a specific operation instruction used to guide the blast furnace operating system to adjust the material. For example, the adjustment instruction may be "add 100 kg of limestone to the blast furnace."

[0152] In this embodiment, the blast furnace operating system is an automated system for controlling and monitoring blast furnace operations, which is capable of receiving adjustment instructions and executing corresponding material adjustments. For example, the blast furnace control system may be an integrated PLC (programmable logic controller) system responsible for real-time monitoring and control of the temperature, pressure, material delivery, etc. in the furnace.

[0153] The working principle and beneficial effects of the above technical solution are: by obtaining the current furnace charge data, inputting the theoretical calculation model to obtain the ideal slag basicity, and calculating the current predicted slag basicity through the prediction model, comparing the ideal and actual basicities, determining the adjustment direction and step size, and then calculating the material adjustment amount, the generated adjustment instructions are passed to the blast furnace operating system to achieve intelligent adjustment of the actual slag basicity, thereby optimizing the ironmaking process, reducing unnecessary material consumption, reducing manual intervention, reducing operational complexity, and improving production efficiency.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent adjustment of blast furnace slag basicity based on big data, characterized in that: include: Step 1: Collect historical production data of blast furnaces, determine key production parameters based on the historical production data, and generate a theoretical calculation model for blast furnace slag basicity; Step 2: Determine the target parameters related to slag basicity in combination with the process principles, screen the related target parameters, obtain strong correlation parameters, and build a prediction model based on the strong correlation parameters; Step 3: Based on the prediction results, the actual data is evaluated to obtain the evaluation results, and the adjustment direction and step length of the actual slag basicity are determined according to the evaluation results; Step 4: Based on the theoretical calculation model and the current charge data, the material adjustment amount is determined in combination with the adjustment direction and adjustment step of the actual slag basicity, and the current actual slag basicity is intelligently adjusted; In step 2, the actual situation of the ironmaking process is obtained, and the process principle is determined based on the actual situation, thereby identifying the first relevant target parameter of slag basicity and the second relevant target parameter fed back by the operator, and calculating the correlation between these two parameters and the slag basicity, constructing a correlation matrix and performing preliminary screening, performing collinearity analysis to merge the parameters, and finally obtaining strongly correlated parameters, and constructing a prediction model based on the strongly correlated parameters; In step 3, the parameter ranges of slag basicity, magnesium-aluminum ratio and molten iron sulfur content are set according to field experience and product quality standards, and state classification is performed accordingly. After obtaining the actual data, the actual state classification corresponding to the actual data is determined, and the actual state classification is compared with the state classification of the predicted results to derive the adjustment direction of the actual slag basicity and calculate the corresponding adjustment step.

2. The method for intelligently adjusting blast furnace slag basicity based on big data according to claim 1, characterized in that: Collect historical production data of blast furnaces, determine key production parameters based on the historical production data, and generate a theoretical calculation model for blast furnace slag basicity, including: Collecting historical production data of a blast furnace, drawing a first box plot using statistical software according to the historical production data, and processing abnormal values ​​in the historical production data according to the first box plot; Identify key production parameters in the processing result according to the first box plot, and draw a second box plot according to the key production parameters to identify parameter distribution characteristics; Draw a heat map for the processing results to determine the connection between each key production parameter and the slag basicity, and determine the influence of different raw material ratios on the slag basicity based on the parameter distribution characteristics and the connection relationship; An initial model is constructed according to the degree of influence combined with a preset basicity adjustment principle. According to the on-site material intake, the current values ​​of key production parameters are input into the initial model to generate a theoretical calculation model for blast furnace slag basicity.

3. The method for intelligently adjusting blast furnace slag basicity based on big data according to claim 1, characterized in that: Combined with the process principles, the target parameters related to slag basicity are determined, and the related target parameters are screened to obtain strong correlation parameters. A prediction model is constructed based on the strong correlation parameters, including: Acquire the actual situation of the ironmaking process, determine the process principle based on the actual situation, determine the first relevant target parameter of slag basicity based on the process principle, collect operator feedback, and obtain the second relevant target parameter of slag basicity; Calculating the correlation between the first relevant target parameter, the second relevant target parameter and the slag basicity, constructing a correlation matrix, and performing a first screening on the first relevant target parameter and the second relevant target parameter according to the correlation matrix to obtain a first screening result; Performing a collinearity analysis on the first relevant target parameter and the second relevant target parameter, and merging the first relevant target parameter and the second relevant target parameter according to the collinearity analysis result to obtain a second screening result; The first screening result and the second screening result are combined to obtain strong correlation parameters, and then a prediction model is constructed.

4. The method for intelligently adjusting blast furnace slag basicity based on big data according to claim 3 is characterized in that: Combining the first screening result with the second screening result to obtain a strong correlation parameter, and then constructing a prediction model, including: Taking blast furnace slag basicity adjustment as the target to be solved, the output type of the prediction model is determined, and the input features are selected from the strongly associated parameters according to the target-feature table; Based on the input features, a deep learning algorithm is used to construct a prediction model for predicting the parameters slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content one smelting cycle in advance.

5. The method for intelligently adjusting blast furnace slag basicity based on big data according to claim 1, characterized in that: Based on the prediction results, the actual data is evaluated to obtain the status evaluation results. According to the status evaluation results, the adjustment direction and adjustment step of the actual slag basicity are determined, including: According to field experience and product quality standards, parameter ranges corresponding to the parameter slag basicity, parameter slag magnesium-aluminum ratio and parameter molten iron sulfur content are set, and the parameter ranges are used to perform state assessment, and then the state classification is determined according to the state assessment results; Acquire actual data, determine the actual state classification corresponding to the actual data according to the state classification, compare the actual state classification with the predicted state classification corresponding to the predicted result, derive the adjustment direction of the actual slag basicity, and calculate the adjustment step size corresponding to the adjustment direction.

6. The method for intelligently adjusting blast furnace slag basicity based on big data according to claim 5 is characterized in that: Calculate the adjustment step size corresponding to the adjustment direction, including: Calculate the adjustment step size: ,in, Indicates the adjustment step of actual slag basicity; K indicates the basic adjustment coefficient; Indicates the actual slag basicity and prediction of slag basicity The influence function of the gap between them on the step length; represents the historical step-size feedback function of the parameter slag basicity; represents the historical adjustment feedback coefficient; Indicates the fluctuation of the parameters slag basicity, slag magnesium-aluminum ratio and molten iron sulfur content in the current time step t; It represents the average value of the parameter slag basicity, parameter slag magnesium-aluminum ratio and parameter molten iron sulfur content in the current time step t; Represents the adjustment coefficient of the coupling factor to the step size; It represents the coupling factor function of the parameter slag basicity, the parameter slag magnesium-aluminum ratio and the parameter molten iron sulfur content; t represents the current time step; T represents the total adjustment period; S represents the stability constant of the smelting process; i represents the parameter slag basicity; j represents the parameter slag magnesium-aluminum ratio; m represents the parameter molten iron sulfur content.

7. The method for intelligently adjusting blast furnace slag basicity based on big data according to claim 6 is characterized in that: Coupling factor functions, including: ,in, It represents the coupling factor function between the parameter slag basicity i and the parameter slag magnesium-aluminum ratio j; represents the actual slag basicity at time step k; Indicates the actual slag basicity The mean value in time step t; It represents the actual slag magnesium-aluminum ratio at time step k; Indicates the actual slag magnesium-aluminum ratio The mean value in time step t; k0 represents the time window size; k represents the time step; ,in, It represents the coupling factor function between the parameter slag basicity i and the parameter molten iron sulfur content m; represents the actual sulfur content of molten iron at time step k; Indicates the actual sulfur content of molten iron The mean value in time step t; ,in, It represents the coupling factor function between the parameter molten iron sulfur content m and the parameter slag magnesium-aluminum ratio j; ,in, The weight coefficient representing the coupling relationship between the parameter slag basicity i and the parameter slag magnesium-aluminum ratio j; The weight coefficient representing the coupling relationship between the parameter slag basicity i and the parameter molten iron sulfur content m; It represents the weight coefficient of the coupling relationship between the parameter molten iron sulfur content m and the parameter slag magnesium-aluminum ratio j.

8. The method for intelligently adjusting blast furnace slag basicity based on big data according to claim 1, characterized in that: Based on the theoretical calculation model and the current charge data, the material adjustment amount is determined in combination with the adjustment direction and step length of the actual slag basicity, and the current actual slag basicity is intelligently adjusted, including: Obtaining current furnace charge data, inputting the current furnace charge data into a theoretical calculation model, and calculating the current ideal slag basicity under ideal conditions under the current furnace charge data according to the theoretical calculation model; Input the current charge data into the prediction model, calculate the current predicted slag basicity under the current charge data according to the prediction model, and then obtain the adjustment direction and adjustment step of the current actual slag basicity; The material adjustment amount is determined based on the adjustment direction and adjustment step of the current ideal slag basicity and the current actual slag basicity, an adjustment instruction is generated according to the material adjustment amount, and the generated adjustment instruction is passed to the blast furnace operating system to make corresponding material adjustments, thereby intelligently adjusting the current actual slag basicity.

Images