A method for predicting blast furnace slag basicity based on big data

Abstract

The present invention provides a method for predicting the basicity of blast furnace slag based on big data, which belongs to the technical field of data processing, and includes: step 1: real-time monitoring and collecting relevant data of smelting of blast furnace based on preset monitoring equipment, processing the relevant data of smelting of blast furnace to obtain relevant data sets of smelting of blast furnace; step 2: analyzing the relevant data sets of smelting of blast furnace, and then constructing a final slag basicity prediction model; step 3: analyzing the relevant data sets of smelting of blast furnace, and constructing a theoretical calculation model of blast furnace slag basicity; step 4: respectively obtaining the output results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model, and then determining the basicity accuracy of each model based on the output results; step 5: determining the final prediction result based on the basicity accuracy of all models and pushing it to the artificial interaction interface. Accurate prediction and intelligent push of blast furnace slag basicity are achieved, and smelting efficiency and process control level are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for predicting blast furnace slag basicity based on big data. Background Art

[0002] In the current blast furnace smelting process, slag basicity is an important indicator that affects smelting efficiency, furnace lining life and final product quality. Traditional basicity prediction methods are mostly based on fixed formulas or empirical values, which are difficult to reflect the complex changes in actual working conditions in real time, resulting in low accuracy and practicality of the prediction results.

[0003] In the prior art, some schemes attempt to combine real-time monitoring data for alkalinity prediction, but the model construction is single and fails to integrate the results of multiple prediction methods, which limits the prediction accuracy and applicability. The defects of the prior art schemes are mainly reflected in: failing to fully utilize multi-model collaborative analysis to improve the accuracy of alkalinity prediction, and failing to comprehensively consider the differences between actual on-site smelting conditions and theoretical calculation models.

[0004] Therefore, the present invention provides a method for predicting blast furnace slag basicity based on big data. Summary of the invention

[0005] The present invention provides a blast furnace slag basicity prediction method based on big data, which is used to construct a multi-model combined basicity prediction method by integrating real-time monitoring data, big data analysis and metallurgical process theory, and optimize the final result by using the model accuracy, thereby solving the problems of low prediction accuracy and poor applicability of traditional methods, realizing accurate prediction and intelligent push of blast furnace slag basicity, and improving smelting efficiency and process control level.

[0006] The present invention provides a method for predicting blast furnace slag basicity based on big data, comprising:

[0007] Step 1: Based on the preset monitoring equipment, real-time monitoring and collection of relevant data of smelting of the blast furnace are performed, and then the smelting relevant data of the blast furnace is processed based on the preset method to obtain the smelting relevant data set of the blast furnace;

[0008] Step 2: Analyze the relevant data sets of blast furnace smelting and then build a final slag basicity prediction model;

[0009] Step 3: Analyze the relevant data sets of blast furnace smelting based on the preset analysis method, and build a theoretical calculation model for blast furnace slag basicity in combination with the actual material consumption on site and the preset metallurgical process theory;

[0010] Step 4: respectively obtain the output results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model, and then determine the basicity accuracy of each model based on the output results;

[0011] Step 5: Determine the final prediction result based on the alkalinity accuracy of all models and push it to the manual interaction interface.

[0012] The present invention provides a method for predicting blast furnace slag basicity based on big data. The smelting-related data of the blast furnace include: raw material data, operation data, smelting status data, slag and iron data, and blast furnace parameter historical data.

[0013] The present invention provides a method for predicting blast furnace slag basicity based on big data, which monitors and collects smelting related data of the blast furnace in real time based on a preset monitoring device, and then processes the smelting related data of the blast furnace based on a preset method to obtain a smelting related data set of the blast furnace, including:

[0014] Deploy preset monitoring equipment at preset key positions of the blast furnace, and monitor and collect smelting-related data of the blast furnace in real time based on the preset monitoring equipment;

[0015] Analyze the smelting-related data of the blast furnace based on the preset method and identify several abnormal data;

[0016] Based on the preset data type-processing method data table, a processing method for abnormal data in the relevant data of smelting of each type of blast furnace is obtained, and the abnormal data is processed;

[0017] Analyze the historical data of the blast furnace parameters to determine the data completeness, and if the data completeness of the historical data of the blast furnace parameters is less than a preset threshold, fill the historical data of the blast furnace parameters based on a preset filling method;

[0018] Based on the smelting related data of the blast furnace after abnormal processing and the historical data of the blast furnace parameters after filling, a relevant data set of blast furnace smelting is constructed.

[0019] The present invention provides a method for predicting blast furnace slag basicity based on big data, which analyzes relevant data sets of blast furnace smelting and then constructs a final slag basicity prediction model, including:

[0020] Acquire several target parameter types that the on-site operators are concerned about, and extract the data of the target parameter types from the relevant data sets of smelting of the blast furnace as the first key parameter;

[0021] Extracting data of a preset parameter type from relevant data sets of smelting of a blast furnace as a second key parameter;

[0022] Performing a first analysis on all first key parameters and second key parameters, thereby eliminating redundant parameters in all first key parameters and second key parameters;

[0023] Integrate the remaining first key parameters and the second key parameters after removing the redundant parameters to generate a first target data set;

[0024] Performing a second analysis on the data set related to smelting of the blast furnace and the data in the target data set, extracting a plurality of third key parameters from the data set related to smelting of the blast furnace based on the second analysis result, and expanding the target data set based on the third key parameters to generate a second target data set;

[0025] Performing a third analysis on the data set related to smelting of the blast furnace and the data in the target data set, extracting a plurality of fourth key parameters from the data set related to smelting of the blast furnace based on the third analysis result, and expanding the target data set based on the fourth key parameters to generate a third target data set;

[0026] Merging the data of the third target data set based on a preset merging method, thereby determining an input feature set of a slag basicity prediction model;

[0027] A slag basicity prediction model is constructed based on an input feature set of the slag basicity prediction model and a preset algorithm;

[0028] The training results of the slag basicity prediction model are evaluated by a preset evaluation method, and the final slag basicity prediction model is determined based on the evaluation results.

[0029] The present invention provides a method for predicting the basicity of blast furnace slag based on big data. The method analyzes the relevant data sets of smelting of the blast furnace based on a preset analysis method, and constructs a theoretical calculation model of the basicity of blast furnace slag in combination with the actual material feeding situation on site and the preset metallurgical process theory, including:

[0030] Analyze the relevant data set of blast furnace smelting based on the preset analysis method, and then extract several key features from the relevant data set of blast furnace smelting based on the preset feature extraction method;

[0031] Analyze the preset metallurgical process theory, and then build several mathematical models based on the actual material consumption on site;

[0032] The key features are taken as the input feature set and all mathematical models are combined to build a theoretical calculation model of blast furnace slag basicity based on the preset model building method.

[0033] The present invention provides a method for predicting blast furnace slag basicity based on big data, respectively obtaining a theoretical calculation model of blast furnace slag basicity and an output result of a final slag basicity prediction model, and then determining the basicity accuracy of each model based on the output result, including:

[0034] Based on the theoretical calculation model of blast furnace slag basicity, theoretical data of slag basicity within a preset smelting cycle are obtained;

[0035] Obtain predicted slag basicity data for a preset smelting cycle based on a final slag basicity prediction model;

[0036] Obtain online measurement data of blast furnace slag and extract the actual basicity data of slag within a preset smelting cycle;

[0037] Determine the verification result of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model based on the theoretical data of slag basicity in the preset smelting cycle, the predicted slag basicity data in the preset smelting cycle and the actual basicity data of slag in the preset smelting cycle;

[0038] The accuracy of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model is determined based on the verification results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model and the preset evaluation method.

[0039] The present invention provides a method for predicting blast furnace slag basicity based on big data, which determines a theoretical calculation model of blast furnace slag basicity and a verification result of a final slag basicity prediction model based on theoretical data of slag basicity in a preset smelting cycle, predicted slag basicity data in a preset smelting cycle, and actual slag basicity data, including:

[0040] Determining a theoretical basicity deviation within a preset smelting cycle based on theoretical basicity data of the slag within a preset smelting cycle and actual basicity data of the slag within the preset smelting cycle;

[0041] Determining a predicted basicity deviation within a preset smelting cycle based on the predicted slag basicity data of the preset smelting cycle and the actual basicity data of the slag within the preset smelting cycle;

[0042] Determine the verification result of the theoretical calculation model of blast furnace slag basicity based on the theoretical basicity deviation within the preset smelting cycle and the theoretical data of slag basicity within the preset smelting cycle;

[0043] The verification result of the final slag basicity prediction model is determined based on the predicted basicity deviation and the predicted slag basicity data of the preset smelting cycle.

[0044] The present invention provides a blast furnace slag basicity prediction method based on big data, which determines the accuracy of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model based on the verification results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model and a preset evaluation method, including:

[0045] The accuracy of the theoretical calculation model of blast furnace slag basicity is determined based on the verification results of the theoretical calculation model of blast furnace slag basicity and the preset evaluation method:

[0046]

[0047] in, is the accuracy of the theoretical calculation model for blast furnace slag basicity, is the theoretical value of slag basicity at the i-th moment in the preset smelting cycle, is the actual basicity value of the slag at the i-th moment in the preset smelting cycle, is the weight corresponding to the actual basicity of the slag at the i-th moment in the preset smelting cycle, and , The verification result of the theoretical calculation model of blast furnace slag basicity is The average value of the actual basicity of the slag during the preset smelting cycle, is the adjustment factor for the actual basicity of the preset slag, and , n1 is the total number of moments in the preset smelting cycle, is the preset adjustment factor of the theoretical calculation model, is the preset deviation adjustment factor, and ;

[0048] The accuracy of the final slag basicity prediction model is determined based on the verification results of the final slag basicity prediction model and the preset evaluation method:

[0049]

[0050] in, is the accuracy of the final slag basicity prediction model, is the predicted slag basicity value at the i-th moment in the preset smelting cycle, is the actual basicity value of the slag at the i-th moment in the preset smelting cycle, is the weight corresponding to the actual basicity of the slag at the i-th moment in the preset smelting cycle, and , is the preset deviation adjustment factor, is the average value of the predicted slag basicity value within the preset smelting cycle, This is the verification result of the final slag basicity prediction model. is the preset adjustment coefficient for predicting slag basicity, Preset adjustment factor for the final slag basicity prediction model.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] By integrating real-time monitoring data, big data analysis and metallurgical process theory, a multi-model combined alkalinity prediction method was constructed, and the final result was optimized using model accuracy. This solved the problems of low prediction accuracy and poor applicability of traditional methods, achieved accurate prediction and intelligent push of blast furnace slag alkalinity, and improved smelting efficiency and process control. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 It is a flow chart of a method for predicting blast furnace slag basicity based on big data provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Embodiment 1:

[0057] The embodiment of the present invention provides a method for predicting blast furnace slag basicity based on big data. Figure 1 As shown, including:

[0058] Step 1: Based on the preset monitoring equipment, real-time monitoring and collection of relevant data of smelting of the blast furnace are performed, and then the smelting relevant data of the blast furnace is processed based on the preset method to obtain the smelting relevant data set of the blast furnace;

[0059] Step 2: Analyze the relevant data sets of blast furnace smelting and then build a final slag basicity prediction model;

[0060] Step 3: Analyze the relevant data sets of blast furnace smelting based on the preset analysis method, and build a theoretical calculation model for blast furnace slag basicity in combination with the actual material consumption on site and the preset metallurgical process theory;

[0061] Step 4: respectively obtain the output results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model, and then determine the basicity accuracy of each model based on the output results;

[0062] Step 5: Determine the final prediction result based on the alkalinity accuracy of all models and push it to the manual interaction interface.

[0063] In this embodiment, the preset monitoring equipment is equipment selected and configured in advance in blast furnace production, and is used to collect relevant production parameters (such as temperature, pressure, composition, etc.) in real time, for example, a sensor installed on the blast furnace, such as a temperature sensor, a gas composition analyzer, or a slag composition monitor;

[0064] In this embodiment, the preset method is a pre-designed data processing and analysis method for standardizing the collected smelting data or generating a preliminary data set, for example, a cleaning algorithm based on big data to remove abnormal data points, or to fill in missing data through interpolation methods.

[0065] In this embodiment, the final slag basicity prediction model is a machine learning or statistical model trained by analyzing historical smelting data and existing production data, and is used to predict the basicity of blast furnace slag. For example, a neural network-based model has inputs including raw material composition, temperature, etc., and outputs a predicted value of slag basicity.

[0066] In this embodiment, the preset analysis method is a specific analysis technique or algorithm pre-selected to mine patterns and regularities in a data set, for example, a principal component analysis (PCA) method to extract key features of smelting data, or to calculate slag basicity based on linear regression;

[0067] In this embodiment, the actual on-site material intake refers to the actual raw material intake at the blast furnace site, including the type, quality and component content of the raw materials, for example, the quality and chemical composition of the sintered ore and coke actually fed, such as CaO and SiO2 content;

[0068] In this embodiment, the preset metallurgical process theory is a formula or method summarized based on theoretical knowledge and experience in the field of metallurgy, which is used to guide the calculation of slag basicity. For example, based on metallurgical theory, slag basicity can be calculated by a formula;

[0069] In this embodiment, the theoretical calculation model of blast furnace slag basicity is a mathematical model for calculating slag basicity based on metallurgical process theory and actual raw material composition and smelting conditions. For example, a model is calculated through feeding data and theoretical formulas to output a theoretical slag basicity value.

[0070] In this embodiment, the basicity accuracy is used to measure the prediction accuracy of the prediction model and the theoretical calculation model for the actual slag basicity, usually expressed as a percentage. If the actual slag basicity is 1.2 and the model predicts 1.18, the accuracy is 98.3%.

[0071] In this embodiment, the final prediction result is the result of integrating the outputs of all models (prediction model and theoretical model), and the most accurate or most suitable slag basicity value for the actual situation is selected as the final output. For example, after comparing multiple models, the final prediction result shows that the slag basicity is 1.25, and it is pushed to the display screen of the blast furnace control room.

[0072] The beneficial effects of the above technical solution are: by integrating real-time monitoring data, big data analysis and metallurgical process theory, a multi-model combined alkalinity prediction method is constructed, and the final result is optimized using model accuracy, which solves the problems of low prediction accuracy and poor applicability of traditional methods, realizes accurate prediction and intelligent push of blast furnace slag alkalinity, and improves smelting efficiency and process control level.

[0073] Embodiment 2:

[0074] The embodiment of the present invention provides a method for predicting blast furnace slag basicity based on big data, and the smelting-related data of the blast furnace include: raw material data, operation data, smelting status data, slag data and blast furnace parameter historical data.

[0075] In this embodiment, the raw material and fuel data are the composition and quality information of the raw materials and fuels used in the blast furnace. For example, raw materials: Fe content in iron ore, SiO 2 Content, Al 2 O 3 Content, fuel: fixed carbon content, volatile matter content, ash ratio of coke;

[0076] In this embodiment, the operation data is to record the key process parameters and adjustment behaviors during the operation of the blast furnace, for example: blast parameters: blast volume, blast temperature, oxygen content, feeding frequency: the time interval for adding raw materials to the blast furnace;

[0077] In this embodiment, the smelting state data reflects the real-time or periodic monitoring data of the blast furnace operation state, for example: temperature: the temperature distribution of each area in the furnace, pressure: the furnace top pressure or the gas flow pressure in the furnace;

[0078] In this embodiment, the slag data is the composition and quality information recorded during the blast furnace tapping and slag tapping process. For example, slag composition: CaO content, SiO 2 Content, MgO content, molten iron composition: carbon content, silicon content, sulfur content;

[0079] In this embodiment, the blast furnace parameter historical data is the operation and production data recorded during the previous blast furnace operation, which is used for analysis and model training. For example: historical alkalinity data: slag alkalinity value measured within a certain period of time, operation parameters: blast historical adjustment records, coal powder injection historical records.

[0080] The beneficial effects of the above technical solution are: by integrating raw materials, operations, smelting status, slag and historical parameter data, a comprehensive blast furnace smelting big data set is constructed, and multi-dimensional data fusion is applied to slag basicity prediction, which improves the comprehensiveness and prediction accuracy of the model and provides reliable support for the optimization of blast furnace smelting process.

[0081] Embodiment 3:

[0082] The embodiment of the present invention provides a method for predicting blast furnace slag basicity based on big data, which monitors and collects smelting related data of the blast furnace in real time based on a preset monitoring device, and then processes the smelting related data of the blast furnace based on a preset method to obtain a smelting related data set of the blast furnace, including:

[0083] Deploy preset monitoring equipment at preset key positions of the blast furnace, and monitor and collect smelting-related data of the blast furnace in real time based on the preset monitoring equipment;

[0084] Analyze the smelting-related data of the blast furnace based on the preset method and identify several abnormal data;

[0085] Based on the preset data type-processing method data table, a processing method for abnormal data in the relevant data of smelting of each type of blast furnace is obtained, and the abnormal data is processed;

[0086] Analyze the historical data of the blast furnace parameters to determine the data completeness, and if the data completeness of the historical data of the blast furnace parameters is less than a preset threshold, fill the historical data of the blast furnace parameters based on a preset filling method;

[0087] Based on the smelting related data of the blast furnace after abnormal processing and the historical data of the blast furnace parameters after filling, a relevant data set of blast furnace smelting is constructed.

[0088] In this embodiment, the preset key positions of the blast furnace are predetermined positions on the blast furnace that are suitable for installing monitoring equipment. These positions can efficiently and comprehensively collect key smelting data, for example, the furnace top: monitor the furnace top pressure and temperature to evaluate the gas flow state in the furnace, the middle of the furnace body: collect the charge descent speed, temperature distribution and gas composition data, the tuyere area: monitor the blast temperature and the state of pulverized coal injection, the iron mouth and slag mouth: record the composition and temperature of molten iron and slag.

[0089] In this embodiment, the preset method is the box plot method, which uses the box plot statistical method to identify abnormal data points through the upper and lower quartiles (Q1 and Q3) and the interquartile range (IQR). For example, the blast furnace temperature data is analyzed: assuming that the data distribution is: [1500, 1520, 1490, 1550, 2000] (unit: ℃), Q1 = 1495, Q3 = 1535, IQR = Q3 - Q1 = 40, abnormal range: less than 1435 or greater than 1595, data point 2000 is marked as abnormal.

[0090] In this embodiment, the method for processing abnormal data is to process abnormal values ​​appearing in the data set, which can be done by deleting the abnormal data or replacing them with reasonable alternative values ​​(such as mean or median). For example, if 2000 is a single abnormal temperature value, it can be directly removed from the data set, and the final data becomes: [1500, 1520, 1490, 1550]. Replace abnormal values: Replace with mean: Calculate the average value of other data 1515, and replace 2000 with 1515. Replace with median: The median of the sorted data is 15101510, and 2000 is replaced with 1510.

[0091] In this embodiment, the filling method is to fill the missing part through a certain algorithm when the historical data is incomplete (such as missing data of certain time periods or not recording certain parameters). If the blast furnace parameter record shows that the furnace top pressure data of a certain hour is missing, the following methods can be used to fill it: Linear interpolation method: fill it according to the average value of the two known data points before and after, for example: the data before and after are 2.0 Pa and 2.4 Pa, and the filling value is 2.2 Pa. Historical mean method: fill it with the historical average value of the parameter under similar working conditions, for example, 2.3 Pa;

[0092] In this embodiment, constructing the relevant data set for smelting is to comprehensively organize the processed abnormal data, the filled historical data and the real-time collected data to form a complete blast furnace smelting data set for subsequent analysis or model training. For example, the data set includes: furnace top pressure: {2.0, 2.2, 2.3, 2.4} (complete data after filling), slag basicity: {1.2,1.3, 1.25} (data after abnormal processing), blast temperature, molten iron composition and other key data. Ultimately, these data provide a unified and standardized input for the basicity prediction model.

[0093] The beneficial effects of the above technical solution are: through real-time monitoring and big data processing, the problem of abnormal and missing blast furnace smelting data is solved, and a complete and accurate data set is constructed. The abnormal processing and data filling methods are adopted to ensure data quality and improve model reliability, providing accurate data foundation and optimization support for blast furnace slag basicity prediction.

[0094] Embodiment 4:

[0095] The embodiment of the present invention provides a method for predicting blast furnace slag basicity based on big data, which analyzes relevant data sets of blast furnace smelting and then constructs a final slag basicity prediction model, including:

[0096] Acquire several target parameter types that the on-site operators are concerned about, and extract the data of the target parameter types from the relevant data sets of smelting of the blast furnace as the first key parameter;

[0097] Extracting data of a preset parameter type from relevant data sets of smelting of a blast furnace as a second key parameter;

[0098] Performing a first analysis on all first key parameters and second key parameters, thereby eliminating redundant parameters in all first key parameters and second key parameters;

[0099] Integrate the remaining first key parameters and the second key parameters after removing the redundant parameters to generate a first target data set;

[0100] Performing a second analysis on the data set related to smelting of the blast furnace and the data in the target data set, extracting a plurality of third key parameters from the data set related to smelting of the blast furnace based on the second analysis result, and expanding the target data set based on the third key parameters to generate a second target data set;

[0101] Performing a third analysis on the data set related to smelting of the blast furnace and the data in the target data set, extracting a plurality of fourth key parameters from the data set related to smelting of the blast furnace based on the third analysis result, and expanding the target data set based on the fourth key parameters to generate a third target data set;

[0102] Merging the data of the third target data set based on a preset merging method, thereby determining an input feature set of a slag basicity prediction model;

[0103] A slag basicity prediction model is constructed based on an input feature set of the slag basicity prediction model and a preset algorithm;

[0104] The training results of the slag basicity prediction model are evaluated by a preset evaluation method, and the final slag basicity prediction model is determined based on the evaluation results.

[0105] In this embodiment, the target parameter type refers to the key parameter type that the on-site operator considers to be very important for the prediction of slag basicity, which is usually related to daily operations, for example, the content of the main components in slag iron (such as CaO, SiO 2 ) and parameters directly related to slag basicity;

[0106] In this embodiment, the first key parameter is the data corresponding to the target parameter type extracted from the data set, for example, the first key parameter: CaO content, SiO 2 Content, sulfur content in molten iron.

[0107] In this embodiment, the preset parameter type is a blast furnace operating parameter closely related to the slag basicity, which is usually determined by field experts or historical research. The second key parameter is data corresponding to the preset parameter type extracted from the data set. For example, the preset parameter type: blast furnace temperature, furnace top pressure, blast composition, coal injection amount, etc. The second key parameter: furnace top temperature: 1500°C, furnace top pressure: 2.3 Pa, coal injection amount: 120 kg / t iron.

[0108] In this embodiment, the first analysis is collinearity analysis, which is used to determine whether there is a high correlation between parameters, so as to eliminate redundant parameters (i.e., parameters that can be expressed by other parameters), for example: parameter 1: blast temperature, parameter 2: furnace middle temperature. If the correlation coefficient of the two is close to 1, it means that they are highly collinear and one of them can be eliminated.

[0109] In this embodiment, the second analysis is a linear analysis for determining a parameter that has a significant linear relationship with the slag basicity. For example, through the linear analysis, it is found that the furnace top pressure is significantly positively correlated with the slag basicity.

[0110] In this embodiment, the third key parameter is a new parameter extracted from the smelting data set by linear analysis, for example, the third key parameter: furnace top pressure (extracted as a key parameter);

[0111] In this embodiment, the second target data set is a new data set generated by adding the third key parameter to the first target data set. For example, the second target data set includes the first key parameter (such as CaO content, SiO 2 content) and the third key parameter (furnace top pressure).

[0112] In this embodiment, the third analysis is a nonlinear analysis, which is used to determine parameters that have a nonlinear relationship with the slag basicity, for example, it is found that the relationship between the blast oxygen content and the slag basicity conforms to a quadratic curve;

[0113] In this embodiment, the fourth key parameter is a new parameter extracted from the smelting data set by nonlinear analysis, for example, blast oxygen content (extracted as a key parameter);

[0114] In this embodiment, the third target data set is a new data set generated by adding the fourth key parameter to the second target data set, for example, including all parameters of the second target data set and the newly added blast oxygen content.

[0115] In this embodiment, the preset merging method is to combine the data obtained from different analyses into a final input feature set according to certain rules, for example, taking all non-redundant parameters, including the first key parameter, the second key parameter, the third key parameter and the fourth key parameter;

[0116] In this embodiment, the input feature set is a set of input variables of the prediction model, which is used to describe the key characteristics of slag basicity. For example, the input feature set: {CaO content, SiO 2 content, furnace top pressure, blast oxygen content}.

[0117] In this embodiment, the preset evaluation method is used to evaluate the performance of the slag basicity prediction model, and a model evaluation index or a cross-validation method is often used. For example: Evaluation index: Mean square error (MSE): measures the average deviation between the predicted value and the actual value, determination coefficient (R²): measures the ability of the model to explain variables, cross-validation method: K-fold cross-validation is used to perform multiple rounds of training and validation on the model to evaluate its stability and generalization ability. Evaluation results: If MSE = 0.02, R² = 0.95, it means that the model has high accuracy and can be used in practical applications.

[0118] The beneficial effects of the above technical solution are: through multiple rounds of data analysis and parameter screening, an innovative blast furnace slag basicity prediction model is constructed and optimized, on-site target parameters and big data mining are combined, the feature set is gradually expanded and refined, the model accuracy and applicability are improved, and the scientificity and stability of basicity prediction are significantly improved, providing efficient data support and optimization solutions for blast furnace operations.

[0119] Embodiment 5:

[0120] The embodiment of the present invention provides a method for predicting the basicity of blast furnace slag based on big data, which analyzes the relevant data sets of smelting of the blast furnace based on a preset analysis method, and constructs a theoretical calculation model of the basicity of blast furnace slag in combination with the actual material feeding situation on site and the preset metallurgical process theory, including:

[0121] Analyze the relevant data set of blast furnace smelting based on the preset analysis method, and then extract several key features from the relevant data set of blast furnace smelting based on the preset feature extraction method;

[0122] Analyze the preset metallurgical process theory, and then build several mathematical models based on the actual material consumption on site;

[0123] The key features are taken as the input feature set and all mathematical models are combined to build a theoretical calculation model of blast furnace slag basicity based on the preset model building method.

[0124] In this embodiment, the mathematical model includes: a slag composition mathematical model, a slag temperature mathematical model, and a blast furnace dynamic operation mathematical model.

[0125] In this embodiment, the preset feature extraction method is a pre-designed rule or algorithm for identifying and extracting characteristic parameters that have a greater impact on slag basicity from relevant data sets of blast furnace smelting. For example: Statistical method: such as using principal component analysis (PCA) to extract several parameters that contribute most to the change in basicity; Data-driven method: using algorithms such as random forest to calculate the importance of each parameter to the prediction of slag basicity, and selecting the top-ranked parameters; Physical law analysis: According to metallurgical process theory, directly select parameters that have a strong correlation with basicity, such as CaO / SiO 2 The ratio of CO / CO in the top gas composition 2 Proportion;

[0126] In this embodiment, the key features are data features that have an important impact on the slag basicity and are obtained by a preset feature extraction method. For example, blast furnace operating conditions: parameters such as furnace top temperature, furnace top pressure, and blast oxygen content that reflect the stability of the smelting process; charge characteristics: CaO content, SiO 2 content, and coke reactivity index, gas composition characteristics: CO content in blast furnace gas, H 2 content, a parameter reflecting the reduction conditions, and the following key features were selected through feature extraction methods (such as PCA analysis): furnace top pressure (2.3 Pa), CaO / SiO 2 ratio (1.2) and CO content (20%), which were used as input features to construct a theoretical calculation model.

[0127] The beneficial effects of the above technical scheme are: extracting the key characteristics of blast furnace smelting through big data analysis, combining the actual material consumption situation with metallurgical process theory, constructing multiple mathematical models and generating alkalinity theoretical calculation models, realizing the deep integration of data-driven and process theory, improving the accuracy and applicability of the model, and providing scientific guidance and optimization support for blast furnace production.

[0128] Embodiment 6:

[0129] The embodiment of the present invention provides a method for predicting blast furnace slag basicity based on big data, respectively obtaining a theoretical calculation model of blast furnace slag basicity and an output result of a final slag basicity prediction model, and then determining the basicity accuracy of each model based on the output result, including:

[0130] Based on the theoretical calculation model of blast furnace slag basicity, theoretical data of slag basicity within a preset smelting cycle are obtained;

[0131] Obtain predicted slag basicity data for a preset smelting cycle based on a final slag basicity prediction model;

[0132] Obtain online measurement data of blast furnace slag and extract the actual basicity data of slag within a preset smelting cycle;

[0133] Determine the verification result of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model based on the theoretical data of slag basicity in the preset smelting cycle, the predicted slag basicity data in the preset smelting cycle and the actual basicity data of slag in the preset smelting cycle;

[0134] The accuracy of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model is determined based on the verification results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model and the preset evaluation method.

[0135] In this embodiment, the theoretical data of slag basicity in the preset smelting cycle is the slag basicity data calculated based on metallurgical process theory and actual parameters through the theoretical calculation model of blast furnace slag basicity in the preset smelting cycle. The theoretical data is derived according to the formula or mathematical model and has not been verified on site. For example: smelting cycle: a blast furnace smelting cycle is 2 hours, theoretical calculation data: based on the charge composition (CaO, SiO 2 content) and blast temperature, the slag basicity calculated by the theoretical model is 1.15 (cycle 1) and 1.18 (cycle 2).

[0136] In this embodiment, the predicted slag basicity data of a preset smelting cycle is the slag basicity data predicted by a final slag basicity prediction model (such as a machine learning model) within the preset smelting cycle based on actual operating conditions data (such as temperature, pressure, gas composition, etc.). For example: smelting cycle: 2 hours, predicted data: use a prediction model (such as random forest) to input blast furnace operating parameters, and the predicted slag basicity obtained is 1.14 (cycle 1), 1.17 (cycle 2).

[0137] In this embodiment, the actual basicity data of the slag is obtained by online measurement equipment (such as a slag component analyzer) or actual laboratory testing, and is the actual basicity value of the slag in a preset smelting cycle. This data is used as a benchmark value to measure the accuracy of the theoretical model and the prediction model. For example: smelting cycle: 2 hours, actual basicity data: detected by online measurement equipment, the actual basicity is 1.16 (cycle 1), 1.18 (cycle 2).

[0138] The beneficial effects of the above technical solution are: by verifying the model output results through online measurement data and real alkalinity data, dynamically evaluating the model accuracy, effectively improving the accuracy and reliability of alkalinity prediction, optimizing the data decision-making ability of the blast furnace smelting process, and helping to achieve efficient and stable smelting quality control.

[0139] Embodiment 7:

[0140] The embodiment of the present invention provides a method for predicting blast furnace slag basicity based on big data, which determines a theoretical calculation model of blast furnace slag basicity and a verification result of a final slag basicity prediction model based on theoretical data of slag basicity in a preset smelting cycle, predicted slag basicity data in a preset smelting cycle, and actual slag basicity data, including:

[0141] Determining a theoretical basicity deviation within a preset smelting cycle based on theoretical basicity data of the slag within a preset smelting cycle and actual basicity data of the slag within the preset smelting cycle;

[0142] Determining a predicted basicity deviation within a preset smelting cycle based on the predicted slag basicity data of the preset smelting cycle and the actual basicity data of the slag within the preset smelting cycle;

[0143] Determine the verification result of the theoretical calculation model of blast furnace slag basicity based on the theoretical basicity deviation within the preset smelting cycle and the theoretical data of slag basicity within the preset smelting cycle;

[0144] The verification result of the final slag basicity prediction model is determined based on the predicted basicity deviation and the predicted slag basicity data of the preset smelting cycle.

[0145] In this embodiment, the theoretical basicity deviation refers to the difference between the basicity data output by the theoretical calculation model of blast furnace slag basicity and the actual measured slag basicity data, reflecting the accuracy of the theoretical model. For example: theoretical basicity data: 1.20, actual basicity data: 1.18

[0146] Theoretical alkalinity deviation: The theoretical data is 0.02 higher than the true value, indicating that the theoretical model is slightly overestimated.

[0147] In this embodiment, the predicted alkalinity deviation refers to the difference between the alkalinity data output by the final slag alkalinity prediction model and the actual slag alkalinity data, reflecting the accuracy of the prediction model. For example: predicted alkalinity data: 1.16, actual alkalinity data: 1.18, predicted alkalinity deviation: the predicted data is 0.02 lower than the actual value, indicating that the prediction model is slightly underestimated.

[0148] In this embodiment, the verification result of the model is based on the overall performance of the theoretical alkalinity deviation or the predicted alkalinity deviation. Combined with the data in multiple cycles, it evaluates whether the theoretical calculation model or prediction model of blast furnace slag alkalinity operates within a reasonable range, and calculates the accuracy or stability of the model. For example: Theoretical model verification result: After analyzing 30 smelting cycles, the average deviation of the theoretical model is ±0.03, and the accuracy is 95%.

[0149] The beneficial effects of the above technical scheme are: by calculating the theoretical basicity deviation and the predicted basicity deviation, integrating the actual slag basicity, theoretical data and predicted data, dynamically verifying and optimizing the theoretical calculation model and the prediction model, the accurate evaluation and correction of the basicity deviation is achieved, the credibility and applicability of the model verification results are improved, and more accurate data support and quality assurance are provided for blast furnace smelting.

[0150] Embodiment 8:

[0151] The embodiment of the present invention provides a method for predicting blast furnace slag basicity based on big data, which determines the accuracy of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model based on the verification results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model and a preset evaluation method, including:

[0152] The accuracy of the theoretical calculation model of blast furnace slag basicity is determined based on the verification results of the theoretical calculation model of blast furnace slag basicity and the preset evaluation method:

[0153]

[0154] in, is the accuracy of the theoretical calculation model for blast furnace slag basicity, is the theoretical value of slag basicity at the i-th moment in the preset smelting cycle, is the actual basicity value of the slag at the i-th moment in the preset smelting cycle, is the weight corresponding to the actual basicity of the slag at the i-th moment in the preset smelting cycle, and , The verification result of the theoretical calculation model of blast furnace slag basicity is The average value of the actual basicity of the slag during the preset smelting cycle, is the adjustment factor for the actual basicity of the preset slag, and , n1 is the total number of moments in the preset smelting cycle, is the preset adjustment factor of the theoretical calculation model, is the preset deviation adjustment factor, and ;

[0155] The accuracy of the final slag basicity prediction model is determined based on the verification results of the final slag basicity prediction model and the preset evaluation method:

[0156]

[0157] in, is the accuracy of the final slag basicity prediction model, is the predicted slag basicity value at the i-th moment in the preset smelting cycle, is the weight corresponding to the actual basicity of the slag at the i-th moment in the preset smelting cycle, and , is the actual basicity value of the slag at the i-th moment in the preset smelting cycle, is the preset deviation adjustment factor, is the average value of the predicted slag basicity value within the preset smelting cycle, This is the verification result of the final slag basicity prediction model. is the preset adjustment coefficient for predicting slag basicity, Preset adjustment factor for the final slag basicity prediction model.

[0158] In this embodiment, the preset adjustment coefficient of the real basicity of the slag is a preset parameter introduced to correct and calibrate the deviation of the real basicity value under specific smelting environment or conditions. This coefficient is used to correct the real basicity deviation caused by measurement error or environmental fluctuation, so that it is more in line with the requirements of the theory or prediction model. For example: real basicity data: the detection value is 1.18, but considering the deviation of the measuring equipment, the adjustment coefficient is 1.02, and the adjusted basicity is: 1.2056 (corrected real basicity);

[0159] In this embodiment, the preset adjustment factor is a parameter used to correct the overall calculation result of the theoretical model or the prediction model, the purpose of which is to improve the matching degree between the model and the actual situation, and is usually determined based on empirical data or historical analysis. For example: the theoretical model calculation value: 1.15, the adjustment factor: 1.05, used to correct the underestimation caused by the simplification of the formula, the adjusted alkalinity: 1.2075;

[0160] In this embodiment, the preset deviation adjustment factor is a fixed value set to balance the errors of certain key parameters in the model, which is used to reduce the deviation impact of the model calculation results and ensure that the results are more accurate. For example: Model calculation deviation: the average deviation between the theoretical value and the true value is ±0.03, the deviation adjustment factor: set to 0.98, which is used to reduce the impact of overestimation. After adjustment, the deviation impact: the calculation result is closer to the true value.

[0161] In this embodiment, the adjustment coefficient for predicting slag basicity is a parameter for correcting the slag basicity value output by the prediction model, the purpose of which is to compensate for the error caused by insufficient model training or fluctuation of input data. For example: predicted value: 1.14, adjustment coefficient: 1.03, used to correct the underestimation caused by input data collection error, the adjusted predicted value: 1.1752;

[0162] In this embodiment, the preset adjustment factor of the final slag basicity prediction model is a parameter for overall optimization of the entire prediction model, which is used to improve the accuracy of the model and reduce the accumulated systematic errors in long-term use. For example: Prediction model output: the average of multiple period prediction values ​​is 1.16, adjustment factor: 1.02, which is used to correct the overall deviation trend. The adjusted prediction result: 1.1832 (closer to the true value).

[0163] The beneficial effects of the above technical scheme are: by introducing the verification results, adjustment coefficients, adjustment factors and deviation adjustment mechanisms of theoretical models and prediction models, combined with big data analysis, the accuracy and stability of blast furnace slag basicity prediction are significantly improved, model errors are dynamically corrected, model performance is optimized, and it adapts to complex changes under different smelting conditions, achieving more accurate and efficient slag basicity management, and effectively ensuring quality control and resource utilization of the smelting process.

[0164] 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 predicting blast furnace slag basicity based on big data, characterized in that: include: Step 1: Based on the preset monitoring equipment, real-time monitoring and collection of relevant data of smelting of the blast furnace are performed, and then the smelting relevant data of the blast furnace is processed based on the preset method to obtain the smelting relevant data set of the blast furnace; Step 2: Analyze the relevant data sets of blast furnace smelting and then build a final slag basicity prediction model; Step 3: Analyze the relevant data sets of blast furnace smelting based on the preset analysis method, and build a theoretical calculation model for blast furnace slag basicity in combination with the actual material consumption on site and the preset metallurgical process theory; Step 4: respectively obtain the output results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model, and then determine the basicity accuracy of each model based on the output results; Step 5: Determine the final prediction result based on the alkalinity accuracy of all models and push it to the manual interaction interface; Wherein, step 4 includes: Based on the theoretical calculation model of blast furnace slag basicity, theoretical data of slag basicity within a preset smelting cycle are obtained; Obtain predicted slag basicity data for a preset smelting cycle based on a final slag basicity prediction model; Obtain online measurement data of blast furnace slag and extract the actual basicity data of slag within a preset smelting cycle; Determine the verification result of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model based on the theoretical data of slag basicity in the preset smelting cycle, the predicted slag basicity data in the preset smelting cycle and the actual basicity data of slag in the preset smelting cycle; The accuracy of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model is determined based on the verification results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model and the preset evaluation method, specifically including: The accuracy of the theoretical calculation model of blast furnace slag basicity is determined based on the verification results of the theoretical calculation model of blast furnace slag basicity and the preset evaluation method: in, is the accuracy of the theoretical calculation model for blast furnace slag basicity, is the theoretical value of slag basicity at the i-th moment in the preset smelting cycle, is the actual basicity value of the slag at the i-th moment in the preset smelting cycle, is the weight corresponding to the actual basicity of the slag at the i-th moment in the preset smelting cycle, and , This is the verification result of the theoretical calculation model of blast furnace slag basicity. The average value of the actual basicity of the slag during the preset smelting cycle, is the adjustment factor for the actual basicity of the preset slag, and , n1 is the total number of moments in the preset smelting cycle, is the preset adjustment factor of the theoretical calculation model, is the preset deviation adjustment factor, and ; The accuracy of the final slag basicity prediction model is determined based on the verification results of the final slag basicity prediction model and the preset evaluation method: in, is the accuracy of the final slag basicity prediction model, is the predicted slag basicity value at the i-th moment in the preset smelting cycle, is the actual basicity value of the slag at the i-th moment in the preset smelting cycle, is the weight corresponding to the actual basicity of the slag at the i-th moment in the preset smelting cycle, and , is the preset deviation adjustment factor, is the average value of the predicted slag basicity value within the preset smelting cycle, This is the verification result of the final slag basicity prediction model. is the preset adjustment coefficient for predicting slag basicity, Preset adjustment factor for the final slag basicity prediction model.

2. The method for predicting blast furnace slag basicity based on big data according to claim 1, characterized in that: The smelting-related data of the blast furnace include: raw material data, operation data, smelting status data, slag data and blast furnace parameter historical data.

3. The method for predicting blast furnace slag basicity based on big data according to claim 1, characterized in that: Based on the preset monitoring equipment, the smelting related data of the blast furnace is monitored and collected in real time, and then the smelting related data of the blast furnace is processed based on the preset method to obtain the smelting related data set of the blast furnace, including: Deploy preset monitoring equipment at preset key positions of the blast furnace, and monitor and collect smelting-related data of the blast furnace in real time based on the preset monitoring equipment; Analyze the smelting-related data of the blast furnace based on the preset method and identify several abnormal data; Based on the preset data type-processing method data table, a processing method for abnormal data in the relevant data of smelting of each type of blast furnace is obtained, and the abnormal data is processed; Analyze the historical data of the blast furnace parameters to determine the data completeness, and if the data completeness of the historical data of the blast furnace parameters is less than a preset threshold, fill the historical data of the blast furnace parameters based on a preset filling method; Based on the smelting related data of the blast furnace after abnormal processing and the historical data of the blast furnace parameters after filling, a relevant data set of blast furnace smelting is constructed.

4. The method for predicting blast furnace slag basicity based on big data according to claim 1, characterized in that: Analyze the relevant data sets of blast furnace smelting and build a final slag basicity prediction model, including: Acquire several target parameter types that the on-site operators are concerned about, and extract the data of the target parameter types from the relevant data sets of smelting of the blast furnace as the first key parameter; Extracting data of a preset parameter type from relevant data sets of smelting of a blast furnace as a second key parameter; Performing a first analysis on all first key parameters and second key parameters, thereby eliminating redundant parameters in all first key parameters and second key parameters; Integrate the remaining first key parameters and the second key parameters after removing the redundant parameters to generate a first target data set; Performing a second analysis on the data set related to smelting of the blast furnace and the data in the target data set, extracting a plurality of third key parameters from the data set related to smelting of the blast furnace based on the second analysis result, and expanding the target data set based on the third key parameters to generate a second target data set; Performing a third analysis on the data set related to smelting of the blast furnace and the data in the target data set, extracting a plurality of fourth key parameters from the data set related to smelting of the blast furnace based on the third analysis result, and expanding the target data set based on the fourth key parameters to generate a third target data set; Merging the data of the third target data set based on a preset merging method, thereby determining an input feature set of a slag basicity prediction model; A slag basicity prediction model is constructed based on an input feature set of the slag basicity prediction model and a preset algorithm; The training results of the slag basicity prediction model are evaluated by a preset evaluation method, and the final slag basicity prediction model is determined based on the evaluation results.

5. The method for predicting blast furnace slag basicity based on big data according to claim 1, characterized in that: Based on the preset analysis method, the relevant data sets of blast furnace smelting are analyzed, and combined with the actual material consumption on site and the preset metallurgical process theory, a theoretical calculation model for blast furnace slag basicity is constructed, including: Analyze the relevant data set of blast furnace smelting based on the preset analysis method, and then extract several key features from the relevant data set of blast furnace smelting based on the preset feature extraction method; Analyze the preset metallurgical process theory, and then build several mathematical models based on the actual material consumption on site; The key features are taken as the input feature set and all mathematical models are combined to build a theoretical calculation model of blast furnace slag basicity based on the preset model building method.

6. The method for predicting blast furnace slag basicity based on big data according to claim 1, characterized in that: The verification results of the theoretical calculation model of blast furnace slag basicity and the final slag basicity prediction model are determined based on the theoretical data of slag basicity within the preset smelting cycle, the predicted slag basicity data of the preset smelting cycle, and the actual basicity data of the slag, including: Determining a theoretical basicity deviation within a preset smelting cycle based on theoretical data of slag basicity within a preset smelting cycle and actual basicity data of slag within the preset smelting cycle; Determining a predicted basicity deviation within a preset smelting cycle based on the predicted slag basicity data of the preset smelting cycle and the actual basicity data of the slag within the preset smelting cycle; Determine the verification result of the theoretical calculation model of blast furnace slag basicity based on the theoretical basicity deviation within the preset smelting cycle and the theoretical data of slag basicity within the preset smelting cycle; The verification result of the final slag basicity prediction model is determined based on the predicted basicity deviation and the predicted slag basicity data of the preset smelting cycle.

Images