A method for predicting blast furnace gas utilization based on process industry data processing

Abstract

The present invention provides a blast furnace gas utilization rate prediction method based on process industry data processing, which relates to the technical field of process data processing, including: collecting various types of blast furnace data, establishing real-time flow processing rules, processing and storing various types of data, cleaning and completing data, and obtaining first processed data; disassembling and integrating the first processed data based on the real-time material discharging situation of the blast furnace, and combining the time lag effect of each data on the gas utilization rate to obtain a first production data set; extracting the monitoring data related to the blast furnace body from the first production data, performing data dimensionality reduction processing, obtaining first dimensionality reduction data, and simultaneously obtaining material distribution related data, establishing a first material distribution model, realizing data derivation, and obtaining first derived data; establishing a gas utilization rate model based on blast furnace characteristics, and combining the first dimensionality reduction data and the first derived data to predict gas utilization rate. This makes the processing of process industry data more accurate, and timely and accurately predicts gas utilization rate.

Description

Technical Field

[0001] The invention relates to the technical field of process data processing, and in particular to a blast furnace gas utilization rate prediction method based on process industry data processing. Background Art

[0002] At present, the blast furnace is an important link in the ironmaking process. Its production process involves many parameters, with the characteristics of long time lag, many influencing factors, and strong parameter coupling. As a result status indicator, the blast furnace gas utilization rate is affected by the raw fuel, furnace conditions, etc., and it also reflects the smooth smelting of the blast furnace and the high or low fuel consumption ratio. In actual production, blast furnace operators usually observe ten temperature measurement changes and gas flow rates at the top of the furnace based on existing identification conditions, and then judge the gas utilization rate based on experience. This traditional method has strong manual dependence, poor judgment stability, and increases the workload of personnel. It does not meet the strategic needs of blast furnace intelligent reform. Therefore, using intelligent algorithm models to predict blast furnace gas utilization will be an inevitable trend in the development of blast furnace intelligence.

[0003] However, the fluctuation of gas utilization rate is affected by many blast furnace production parameters. Existing methods or manual experience are highly dependent, resulting in labor intensity and untimely production adjustments; or the parameters are not considered comprehensively, resulting in weak model generalization ability and reduced universality, which will indirectly lead to increased production costs and reduced efficiency.

[0004] Therefore, the present invention provides a method for predicting blast furnace gas utilization based on process industry data processing. Summary of the invention

[0005] The present invention provides a blast furnace gas utilization rate prediction method based on process industry data processing, which is used to collect, process and analyze data generated in the blast furnace production process, so that the processing of process industry data is more accurate, and the gas utilization rate of the target blast furnace can be predicted more timely and accurately, so as to improve energy utilization and reduce enterprise energy consumption.

[0006] The present invention provides a blast furnace gas utilization rate prediction method based on process industry data processing, comprising:

[0007] Step 1: Collect various data of the target blast furnace, and establish real-time flow processing rules of the target blast furnace based on computer and database technology, so as to process and store various data of the target blast furnace, clean and complete the data, and obtain first processed data;

[0008] Step 2: Based on the real-time unloading situation of the target blast furnace, the first processed data is decomposed and integrated, and the first production data set of the target blast furnace is obtained by combining the time lag effect of each data on the gas utilization rate in the decomposition and integration results;

[0009] Step 3: extracting the monitoring data related to the target blast furnace body from the first production data set to obtain the first monitoring data, and performing data dimensionality reduction processing to obtain the first dimensionality reduction data, and at the same time obtaining the distribution-related data of the target blast furnace, thereby establishing the first distribution model of the target blast furnace, realizing data derivation, and obtaining the first derived data;

[0010] Step 4: Establish a corresponding gas utilization model based on the blast furnace characteristics of the target blast furnace and the machine learning algorithm, and predict the gas utilization of the target blast furnace based on the first dimensionality reduction data and the first derivative data.

[0011] According to the present invention, various data of the target blast furnace are collected, and real-time flow processing rules of the target blast furnace are established based on computer and database technology, so that various data of the target blast furnace are processed and stored, and data cleaning and completion are performed to obtain first processed data, including:

[0012] Step 11: various data are collected in real time based on various sensors of the target blast furnace to obtain first collected data;

[0013] Step 12: Determine a real-time data processing flow that matches the target blast furnace based on computer and database technology combined with the blast furnace characteristics of the target blast furnace, and determine a real-time data flow processing rule based on the data type of the blast furnace data and the first collected data;

[0014] Step 13: Perform data conversion and data storage on the first collected data based on the real-time flow processing rule to obtain second collected data;

[0015] Step 14: Clean and complete the second collected data to obtain the first processed data of the target blast furnace.

[0016] According to the present invention, the first processing data is decomposed and integrated based on the real-time material unloading situation of the target blast furnace, and the first production data set of the target blast furnace is obtained by combining the time lag effect of each data on the gas utilization rate in the decomposition and integration results, including:

[0017] Step 21: Based on the real-time material unloading situation of the target blast furnace, the first processed data is decomposed, so that the charging data in the decomposed first processed data corresponds to the chemical composition and physical properties of the raw fuel data;

[0018] Step 22: classify and integrate the remaining processed data of the target blast furnace except the charging data in the first processed data according to different data sources, so as to construct an initial production data set of the target blast furnace in combination with the processing result of the charging data;

[0019] Step 23: Analyze whether each data in the initial production data set has a time lag effect on the gas utilization rate, and determine the time lag time of the data with time lag effect based on a preset cross-correlation analysis method, and adjust the initial production data set according to the time lag time to obtain the first production data set.

[0020] According to the present invention, the method for analyzing whether each data in the initial production data set has a time lag effect on the gas utilization rate, and determining the time lag time of the data with time lag effect based on a preset cross-correlation analysis method, and adjusting the initial production data set according to the time lag time to obtain a first production data set includes:

[0021] Step 231: obtaining the data type and data characteristics of each data in the initial production data set, thereby extracting the first historical data of the corresponding data type in the target blast furnace;

[0022] Step 232: obtaining a first data reference based on the data characteristics of the first historical data and the data type related to the gas utilization rate, and selecting a corresponding first analysis method based on the first data reference;

[0023] Step 233: determining whether there is a time lag effect on the gas utilization rate during the historical working process of the first historical data based on the first analysis method;

[0024] If there is a time lag effect, the current data in the initial production data set is used as the first data;

[0025] If there is no time lag effect, the current data in the initial production data set is used as the second data;

[0026] Step 234: obtaining a first data set based on each first data integration of the initial production data set, and obtaining a second data set based on each second data integration of the initial production data set;

[0027] Step 235: performing data processing on each first data in the first data set based on a preset cross-correlation analysis method, so as to determine a cross-correlation coefficient Ti corresponding to each first data when the lag time is i;

[0028] ; Where Ti is the cross-correlation coefficient corresponding to the lag time i, is the observed value of the time series h at time ti, is the observed value of the time series s at time t, is the mean of the h time series, is the mean of the time series s, n is the number of observations in the time series, is the error affected by the sequence stability of the h time series, is the error affected by the sequence stability of the s time series, The error is caused by noise;

[0029] Step 236: comparing the cross-correlation coefficients corresponding to each lag time, and extracting the lag time corresponding to the maximum cross-correlation coefficient as the first lag time of the current first data;

[0030] Step 237: adjusting the relative time series of the current first data corresponding to the first data set based on the first time lag time, so as to achieve consistency of the first data in the time series, thereby obtaining a first adjusted data set;

[0031] Step 238: Integrate the first adjustment data set with the second data set to obtain a first production data set of the target blast furnace.

[0032] According to the present invention, the monitoring data related to the target blast furnace body is extracted from the first production data set to obtain the first monitoring data, and the data dimension reduction is performed to obtain the first dimension reduction data, and the distribution related data of the target blast furnace is obtained at the same time, so as to establish the first distribution model of the target blast furnace, realize data derivation, and obtain the first derived data, including:

[0033] Step 31: extracting data related to the target blast furnace body from the first production data set to obtain first monitoring data;

[0034] Step 32: performing data dimensionality reduction processing on the first monitoring data based on a preset dimensionality reduction method to obtain first dimensionality reduction data;

[0035] Step 33: acquiring real-time charge distribution data of the target blast furnace, and extracting a suitable charge distribution data model from a charge distribution model database to construct a first charge distribution model in combination with charge distribution characteristics of the target blast furnace during charge distribution;

[0036] Among them, real-time material distribution related data includes material batch feeding data and material distribution matrix data;

[0037] Step 34: Input the real-time charge distribution related data into the first charge distribution model, so as to obtain the first derivative data of the target blast furnace.

[0038] According to the present invention, the real-time charge distribution related data of the target blast furnace is obtained, and combined with the charge distribution characteristics of the target blast furnace during blast furnace charge distribution, a suitable charge distribution data model is extracted from the charge distribution model database to construct a first charge distribution model, including:

[0039] Step 331: acquiring real-time charge distribution related data of the target blast furnace, and acquiring charge characteristic parameters of the target blast furnace for blast furnace charge distribution;

[0040] Step 332: obtaining the information difference between the real-time material surface information and the standard material surface information, thereby constructing a blast furnace charging matrix of the target blast furnace;

[0041] Step 333: selecting an initial blast furnace charge distribution model matching the target blast furnace from a charge distribution model database based on the charge characteristic parameters and charge information of the target blast furnace;

[0042] Step 334: inputting the blast furnace charge distribution matrix into the initial blast furnace charge distribution model, thereby obtaining second charge surface information;

[0043] Step 335: Obtain the charge distribution optimization target of the target blast furnace, thereby comparing the second charge surface information with the charge distribution optimization target, and adjusting the parameters of the blast furnace charge distribution matrix based on the comparison result, thereby verifying the initial blast furnace charge distribution model and obtaining a first charge distribution model.

[0044] According to the present invention, a corresponding gas utilization model is established based on the blast furnace characteristics of the target blast furnace and the machine learning algorithm, and the gas utilization rate of the target blast furnace is predicted based on the first dimensionality reduction data and the first derivative data, including:

[0045] Step 41: Selecting a machine learning algorithm based on the blast furnace characteristics of the target blast furnace to construct a gas utilization model of the target blast furnace;

[0046] Step 42: inputting the first dimension reduction data and the first derivative data into the gas utilization rate model, thereby predicting the gas utilization rate of the target blast furnace based on each gas utilization rate model to obtain a first prediction result;

[0047] Step 43: Based on the first prediction result, data optimization is performed on corresponding data in the first production data set of the target blast furnace. At the same time, based on the first prediction result, matrix optimization is performed on the data matrix of the distribution matrix data in the distribution-related data of the target blast furnace, so as to improve the gas utilization rate of the target blast furnace.

[0048] According to the present invention, data optimization is performed on corresponding data in the first production data set of the target blast furnace based on the first prediction result, and at the same time, matrix optimization is performed on the data matrix of the distribution matrix data in the distribution-related data of the target blast furnace based on the first prediction result, thereby improving the gas utilization rate of the target blast furnace, including:

[0049] Step 431: obtaining an expected target of real-time gas utilization rate of a target blast furnace, and comparing the expected target of real-time gas utilization rate with the first prediction result, thereby obtaining a first comparison result;

[0050] Step 432: Based on the first comparison result, identifying data whose impact on the gas utilization rate exceeds a preset minimum impact, as the first key data of the target blast furnace;

[0051] Step 433: extracting the comparison result corresponding to the first key data from the first comparison result to obtain a second comparison result;

[0052] Step 434: extracting a data optimization strategy corresponding to the second comparison result from the prediction-optimization database, and performing data optimization on the first key data, thereby improving the gas utilization rate of the target blast furnace;

[0053] Step 435: Based on the first comparison result and in combination with the first charging model of the target blast furnace, matrix optimization is performed on the data matrix of the charging matrix data in the charging-related data of the target blast furnace, thereby improving the gas utilization rate of the target blast furnace.

[0054] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a blast furnace gas utilization prediction method based on process industry data processing, which collects, processes and analyzes data generated in the blast furnace production process, thereby making the processing of process industry data more accurate, and being able to more timely and accurately predict the gas utilization rate of the target blast furnace, so as to improve energy utilization and reduce enterprise energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] 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.

[0056] Figure 1 The present invention provides a flowchart of a method for predicting blast furnace gas utilization rate based on process industry data processing. DETAILED DESCRIPTION

[0057] 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.

[0058] Embodiment 1:

[0059] The embodiment of the present invention provides a method for predicting blast furnace gas utilization rate based on process industry data processing. Figure 1 As shown, including:

[0060] Step 1: Collect various data of the target blast furnace, and establish real-time flow processing rules of the target blast furnace based on computer and database technology, so as to process and store various data of the target blast furnace, clean and complete the data, and obtain first processed data;

[0061] Step 2: Based on the real-time unloading situation of the target blast furnace, the first processed data is decomposed and integrated, and the first production data set of the target blast furnace is obtained by combining the time lag effect of each data on the gas utilization rate in the decomposition and integration results;

[0062] Step 3: extracting the monitoring data related to the target blast furnace body from the first production data set to obtain the first monitoring data, and performing data dimensionality reduction processing to obtain the first dimensionality reduction data, and at the same time obtaining the distribution-related data of the target blast furnace, thereby establishing the first distribution model of the target blast furnace, realizing data derivation, and obtaining the first derived data;

[0063] Step 4: Establish a corresponding gas utilization model based on the blast furnace characteristics of the target blast furnace and the machine learning algorithm, and predict the gas utilization of the target blast furnace based on the first dimensionality reduction data and the first derivative data.

[0064] In this embodiment, the target blast furnace refers to a specific blast furnace equipment, which is an important link in the steel production process and is used to reduce iron ore into iron.

[0065] In this embodiment, computer and database technology refers to the technology of processing, storing and managing data using computer hardware, software and database management system. In blast furnace data processing, these technologies are used to establish real-time flow processing rules to ensure the accuracy and timeliness of data.

[0066] In this embodiment, the real-time flow processing rules are a set of rules and processes for processing real-time data, ensuring that the data can be collected, processed, stored and transmitted in a predetermined manner. In blast furnace data processing, these rules may include the format, transmission method, storage location, etc. of the data.

[0067] In this embodiment, data cleaning refers to the process of checking for errors, anomalies or duplicate values ​​in the data and correcting or deleting them.

[0068] In this embodiment, data completion refers to supplementing missing data to ensure the integrity and accuracy of the data.

[0069] In this embodiment, the first processed data refers to the data obtained after data cleaning and completion, which is the basis for subsequent analysis and modeling.

[0070] In this embodiment, the real-time material feeding situation refers to the situation and implementation time of the raw materials (such as iron ore, coke, etc.) being added to the blast furnace in real time during the production process of the blast furnace. It reflects the production status and efficiency of the blast furnace.

[0071] In this embodiment, data decomposition refers to breaking down complex data sets into smaller, more easily analyzed parts.

[0072] In this embodiment, data integration refers to combining data from multiple data sources into a unified data set for comprehensive analysis.

[0073] In this embodiment, the gas utilization rate refers to the proportion of the gas generated in the blast furnace that is effectively utilized, and is an important indicator for measuring the energy efficiency of the blast furnace.

[0074] In this embodiment, the time lag effect means that there is a certain delay in the impact of a certain parameter or event on the current or future gas utilization rate. In blast furnace production, changes in certain operations or parameters may not be immediately reflected in the gas utilization rate, but have a certain lag effect.

[0075] In this embodiment, the first production data set is a data set obtained based on the real-time material unloading situation of the target blast furnace, data decomposition and integration, and time lag impact analysis.

[0076] In this embodiment, the monitoring data related to the target blast furnace body refers to the monitoring data that directly reflects the operating status of the blast furnace body, such as the blast furnace temperature, pressure, gas flow rate, etc.

[0077] In this embodiment, data dimension reduction is a technique for reducing the number of features in a data set, aiming to reduce the complexity of the data while retaining the main information of the data. In blast furnace data analysis, data dimension reduction can help us find the key factors affecting gas utilization more quickly.

[0078] In this embodiment, the first dimension reduction data refers to a data set obtained after data dimension reduction processing.

[0079] In this embodiment, the distribution-related data refers to the data of the raw material batch added to the blast furnace each time, including the raw material type, quantity, time, distribution matrix, etc.

[0080] In this embodiment, the first charge distribution model is a model established based on charge distribution related data of the target blast furnace, and is used to describe the charge distribution situation inside the blast furnace.

[0081] In this embodiment, data derivation refers to the process of generating new and valuable information or data from raw data through data analysis, modeling or conversion. In a blast furnace gas utilization rate prediction method based on process industry data processing, data derivation is achieved by establishing a distribution data model.

[0082] In this embodiment, the first derived data is a result obtained by deriving data based on the first cloth model.

[0083] In this embodiment, the blast furnace characteristics refer to the properties of the target blast furnace in terms of structure, operation, etc.

[0084] In this embodiment, the gas utilization model is constructed based on a machine learning algorithm and is a model for predicting a target blast furnace gas utilization rate. The model can predict the gas utilization rate of a blast furnace based on input data (such as reduced dimensionality data, derived data, etc.).

[0085] The beneficial effect of the above technical solution is: by collecting, processing and analyzing the data generated in the blast furnace production process, the processing of process industry data is more accurate, and the gas utilization rate of the target blast furnace can be predicted more timely and accurately, so as to improve energy utilization and reduce enterprise energy consumption.

[0086] Embodiment 2:

[0087] Based on Example 1, the first processed data is obtained, including:

[0088] Step 11: various data are collected in real time based on various sensors of the target blast furnace to obtain first collected data;

[0089] Step 12: Determine a real-time data processing flow that matches the target blast furnace based on computer and database technology combined with the blast furnace characteristics of the target blast furnace, and determine a real-time data flow processing rule based on the data type of the blast furnace data and the first collected data;

[0090] Step 13: Perform data conversion and data storage on the first collected data based on the real-time flow processing rule to obtain second collected data;

[0091] Step 14: Clean and complete the second collected data to obtain the first processed data of the target blast furnace.

[0092] In this embodiment, the target blast furnace refers to a specific blast furnace equipment, which is an important link in the steel production process and is used to reduce iron ore into iron.

[0093] In this embodiment, the first collected data is raw data collected in real time by the sensor and is the basis for data processing and analysis.

[0094] In this embodiment, the blast furnace characteristics refer to the structure, operating conditions, production capacity and other characteristics of the blast furnace.

[0095] In this embodiment, computer and database technology refers to the technology of processing, storing and managing data using computer hardware, software and database management system. In blast furnace data processing, these technologies are used to establish real-time flow processing rules to ensure the accuracy and timeliness of data.

[0096] In this embodiment, the real-time flow processing rules are a set of rules and processes for processing real-time data, ensuring that the data can be collected, processed, stored and transmitted in a predetermined manner. In blast furnace data processing, these rules may include the format, transmission method, storage location, etc. of the data.

[0097] In this embodiment, data conversion is the process of converting data from one format or type to another format or type. In blast furnace data processing, data conversion may include converting sensor data into a storable format, or converting data into a form suitable for analysis.

[0098] In this embodiment, data storage refers to storing data in a computer system for analysis and use. In blast furnace data processing, data storage usually involves storing processed data in a database.

[0099] In this embodiment, the second collected data is data obtained after data conversion and storage.

[0100] In this embodiment, data cleaning refers to the process of checking for errors, anomalies or duplicate values ​​in the data and correcting or deleting them.

[0101] In this embodiment, data completion refers to supplementing missing data to ensure the integrity and accuracy of the data.

[0102] In this embodiment, the first processed data refers to the data obtained after data cleaning and completion, which is the basis for subsequent analysis and modeling.

[0103] The beneficial effect of the above technical solution is: by collecting and processing the data generated during the blast furnace production process, more accurate data analysis can be performed, making the processing of process industry data more accurate, and being able to predict the gas utilization rate of the target blast furnace more timely and accurately.

[0104] Embodiment 3:

[0105] Based on Example 2, the first production data set of the target blast furnace is obtained by combining the time lag effect of each data in the disassembly and integration results on the gas utilization rate, including:

[0106] Step 21: Based on the real-time material unloading situation of the target blast furnace, the first processed data is decomposed, so that the charging data in the decomposed first processed data corresponds to the chemical composition and physical properties of the raw fuel data;

[0107] Step 22: classify and integrate the remaining processed data of the target blast furnace except the charging data in the first processed data according to different data sources, so as to construct an initial production data set of the target blast furnace in combination with the processing result of the charging data;

[0108] Step 23: Analyze whether each data in the initial production data set has a time lag effect on the gas utilization rate, and determine the time lag time of the data with time lag effect based on a preset cross-correlation analysis method, and adjust the initial production data set according to the time lag time to obtain the first production data set.

[0109] In this embodiment, the real-time material feeding situation refers to the situation in which raw materials (such as iron ore, coke, etc.) are added to the blast furnace in real time during the production process of the blast furnace, which reflects the production status and efficiency of the blast furnace.

[0110] In this embodiment, the data decomposition is to match the charge data in the first processing data with the chemical composition and physical properties of the raw fuel data.

[0111] In this embodiment, the charge data refers to the raw material data entering the blast furnace, including information such as the type, quantity, chemical composition, and physical properties of the raw materials.

[0112] In this embodiment, the raw material and fuel data refer to the data of the raw materials and fuels used in the blast furnace, and these data usually include key information such as chemical composition and physical properties.

[0113] In this embodiment, chemical composition refers to the various chemical elements and their proportions contained in the raw materials or fuels; physical properties refer to the physical characteristics of the raw materials or fuels, such as density, hardness, melting point, etc.

[0114] In this embodiment, classification integration is to classify data according to certain standards or rules, and integrate the classified data. Here, classification integration is to classify the remaining processed data except the charge data in the first processed data according to different data sources, and construct the initial production data set of the target blast furnace in combination with the processing results of the charge data.

[0115] In this embodiment, the initial production data set is a data set obtained after data decomposition and classification integration, and includes various data generated by the target blast furnace during the production process.

[0116] In this embodiment, the gas utilization rate refers to the proportion of the gas generated in the blast furnace that is effectively utilized, and is an important indicator for measuring the energy efficiency of the blast furnace.

[0117] In this embodiment, data integration refers to combining data from multiple data sources into a unified data set for comprehensive analysis.

[0118] In this embodiment, the time lag effect means that there is a certain delay in the impact of a certain parameter or event on the current or future gas utilization rate. In blast furnace production, changes in certain operations or parameters may not be immediately reflected in the gas utilization rate, but have a certain lag effect.

[0119] In this embodiment, the first production data set is a data set obtained based on the real-time material unloading situation of the target blast furnace, data decomposition and integration, and time lag impact analysis.

[0120] The beneficial effect of the above technical solution is: by conducting more accurate data integration and analysis on the data generated during the blast furnace production process, the processing of process industry data is made more accurate, and the gas utilization rate of the target blast furnace can be predicted more timely and accurately to improve energy utilization.

[0121] Embodiment 4:

[0122] Based on Example 3, a first production data set is obtained, including:

[0123] Step 231: obtaining the data type and data characteristics of each data in the initial production data set, thereby extracting the first historical data of the corresponding data type in the target blast furnace;

[0124] Step 232: obtaining a first data reference based on the data characteristics of the first historical data and the data type related to the gas utilization rate, and selecting a corresponding first analysis method based on the first data reference;

[0125] Step 233: determining whether there is a time lag effect on the gas utilization rate during the historical working process of the first historical data based on the first analysis method;

[0126] If there is a time lag effect, the current data in the initial production data set is used as the first data;

[0127] If there is no time lag effect, the current data in the initial production data set is used as the second data;

[0128] Step 234: obtaining a first data set based on each first data integration of the initial production data set, and obtaining a second data set based on each second data integration of the initial production data set;

[0129] Step 235: performing data processing on each first data in the first data set based on a preset cross-correlation analysis method, so as to determine a cross-correlation coefficient Ti corresponding to each first data when the lag time is i;

[0130] ; Where Ti is the cross-correlation coefficient corresponding to the lag time i, is the observed value of the time series h at time ti, is the observed value of the time series s at time t, is the mean of the h time series, is the mean of the time series s, n is the number of observations in the time series, is the error affected by the sequence stability of the h time series, is the error affected by the sequence stability of the s time series, The error is caused by noise;

[0131] Step 236: comparing the cross-correlation coefficients corresponding to each lag time, and extracting the lag time corresponding to the maximum cross-correlation coefficient as the first lag time of the current first data;

[0132] Step 237: adjusting the relative time series of the current first data corresponding to the first data set based on the first time lag time, so as to achieve consistency of the first data in the time series, thereby obtaining a first adjusted data set;

[0133] Step 238: Integrate the first adjustment data set with the second data set to obtain a first production data set of the target blast furnace.

[0134] In this embodiment, the initial production data set refers to a data set that contains various data generated during the production process of the target blast furnace after preliminary processing and analysis.

[0135] In this embodiment, data type refers to the type or format of data, such as numbers, text, etc.; data characteristics refer to specific properties or characteristics exhibited by data, such as distribution, range, and change trend of data.

[0136] In this embodiment, the first historical data refers to historical data corresponding to the current data type in the target blast furnace.

[0137] In this embodiment, the first data reference refers to a data standard or basis for guiding subsequent analysis, which is obtained based on the data characteristics of the first historical data and the data type related to the gas utilization rate.

[0138] In this embodiment, the first analysis method refers to a method selected based on the first data reference and used to analyze the impact of the first historical data on the gas utilization rate.

[0139] In this embodiment, the time lag effect means that there is a certain delay in the impact of a certain parameter or event on the current or future gas utilization rate. In blast furnace production, changes in certain operations or parameters may not be immediately reflected in the gas utilization rate, but have a certain lag effect.

[0140] In this embodiment, the first data refers to data affected by time lag, and the second data refers to data not affected by time lag, which are determined according to the analysis result of the first historical data by the first analysis method.

[0141] In this embodiment, the first data set is a set consisting of all first data, and the second data set is a set consisting of all second data.

[0142] In this embodiment, the cross-correlation analysis method is used to analyze the correlation between each first data in the first data set and the gas utilization rate at different lag times.

[0143] In this embodiment, the cross-correlation coefficient Ti refers to the cross-correlation coefficient between a certain first data in the first data set and the gas utilization rate when the lag time is i, and the cross-correlation coefficient reflects the degree of correlation between the two.

[0144] In this embodiment, the first time lag time refers to the time delay required for a certain first data to affect the gas utilization rate, that is, the lag time corresponding to the maximum cross-correlation coefficient.

[0145] In this embodiment, the consistency of the time series refers to the consistency of the relative position or order between the data in the time series. The consistency of the first data in the time series is achieved by adjusting the relative time series of the data in the first data set. When the cross-correlation coefficient is a positive value, it indicates that the two time series are positively correlated; when the cross-correlation coefficient is a negative value, it indicates that the two time series are negatively correlated; when the cross-correlation coefficient is 0, it indicates that the two time series are unrelated.

[0146] In this embodiment, the first adjusted data set refers to a first data set adjusted based on the first time lag, and the data in this set remains consistent in time series.

[0147] In this embodiment, the first production data set is a data set obtained based on the real-time material unloading situation of the target blast furnace, data decomposition and integration, and time lag impact analysis.

[0148] In this embodiment, the beneficial effect of the above technical solution is: by performing more accurate data integration and analysis on the data generated during the blast furnace production process, the processing of process industry data is made more accurate, and the gas utilization rate of the target blast furnace can be predicted more timely and accurately to improve energy utilization.

[0149] Embodiment 5:

[0150] Based on Example 3, a first distribution model of a target blast furnace is established to achieve data derivation and obtain first derived data, including:

[0151] Step 31: extracting data related to the target blast furnace body from the first production data set to obtain first monitoring data;

[0152] Step 32: performing data dimensionality reduction processing on the first monitoring data based on a preset dimensionality reduction method to obtain first dimensionality reduction data;

[0153] Step 33: acquiring real-time charge distribution data of the target blast furnace, and extracting a suitable charge distribution data model from a charge distribution model database to construct a first charge distribution model in combination with charge distribution characteristics of the target blast furnace during charge distribution;

[0154] Among them, real-time material distribution related data includes material batch feeding data and material distribution matrix data;

[0155] Step 34: Input the real-time charge distribution related data into the first charge distribution model, so as to obtain the first derivative data of the target blast furnace.

[0156] In this embodiment, the monitoring data related to the target blast furnace body refers to the monitoring data that directly reflects the operating status of the blast furnace body, such as the blast furnace temperature, pressure, gas flow rate, etc.

[0157] In this embodiment, data dimension reduction is a technique for reducing the number of features in a data set, aiming to reduce the complexity of the data while retaining the main information of the data. In blast furnace data analysis, data dimension reduction can help us find the key factors affecting gas utilization more quickly.

[0158] In this embodiment, the first monitoring data is extracted from the first production data set and is data related to the target blast furnace body.

[0159] In this embodiment, the preset dimensionality reduction method refers to one or more data dimensionality reduction methods that have been determined before the data dimensionality reduction process is performed.

[0160] In this embodiment, the first dimension reduction data refers to a data set obtained after data dimension reduction processing.

[0161] In this embodiment, the distribution-related data refers to the data of the raw material batch added to the blast furnace each time, including the raw material type, quantity, time, distribution matrix, etc.

[0162] In this embodiment, the burden distribution characteristics of the blast furnace refer to specific rules or features exhibited during the burden distribution process of the blast furnace.

[0163] In this embodiment, the burden distribution model database is a database storing a variety of burden distribution data models. These models are used to describe and predict various phenomena and laws in the blast furnace burden distribution process.

[0164] In this embodiment, the first charge distribution model is established based on the first dimension reduction data and is a data model used to describe the charge distribution situation of the blast furnace.

[0165] In this embodiment, the prediction accuracy of the blast furnace gas utilization rate prediction refers to the degree of consistency between the prediction result and the actual situation when predicting the blast furnace gas utilization rate. The higher the prediction accuracy, the more reliable the prediction result.

[0166] In this embodiment, the minimum impact refers to the degree of influence of the data feature that has the least impact on the blast furnace gas utilization prediction result when extracting features. Only data features with a higher impact than the minimum impact will be extracted and used for model construction.

[0167] In this embodiment, data derivation refers to the process of generating new and valuable information or data from raw data through data analysis, modeling or conversion. In a blast furnace gas utilization rate prediction method based on process industry data processing, data derivation is achieved by establishing a distribution data model.

[0168] In this embodiment, the first derived data is the data output by the model after the data with a higher than minimum impact degree in the real-time charge distribution related data is input as input parameters into the first charge distribution model. These data are calculated based on the real-time charge distribution related data and the charge distribution model, and reflect the key information in the blast furnace charge distribution process.

[0169] The beneficial effect of the above technical solution is: by performing dimensionality reduction processing on the data generated in the blast furnace production process and conducting comprehensive analysis based on the data derivative results of the first distribution model, the gas utilization rate of the target blast furnace can be predicted more timely and accurately, so as to improve energy utilization and reduce enterprise energy consumption.

[0170] Embodiment 6:

[0171] Based on Example 5, a first cloth model is constructed, including:

[0172] Step 331: acquiring real-time charge distribution related data of the target blast furnace, and acquiring charge characteristic parameters of the target blast furnace for blast furnace charge distribution;

[0173] Step 332: obtaining the information difference between the real-time material surface information and the standard material surface information, thereby constructing a blast furnace charging matrix of the target blast furnace;

[0174] Step 333: selecting an initial blast furnace charge distribution model matching the target blast furnace from a charge distribution model database based on the charge characteristic parameters and charge information of the target blast furnace;

[0175] Step 334: inputting the blast furnace charge distribution matrix into the initial blast furnace charge distribution model, thereby obtaining second charge surface information;

[0176] Step 335: Obtain the charge distribution optimization target of the target blast furnace, thereby comparing the second charge surface information with the charge distribution optimization target, and adjusting the parameters of the blast furnace charge distribution matrix based on the comparison result, thereby verifying the initial blast furnace charge distribution model and obtaining a first charge distribution model.

[0177] In this embodiment, the charge characteristic parameters refer to the physical and chemical characteristic parameters of the charge in the blast furnace, such as particle size, density, chemical composition, etc.

[0178] In this embodiment, the real-time material surface information refers to the real-time material surface distribution inside the blast furnace, which is usually obtained through sensors or monitoring equipment inside the blast furnace.

[0179] In this embodiment, the standard material surface information refers to the material surface distribution expected to be achieved during the blast furnace charging process, which is usually determined based on the operating conditions and operating experience of the blast furnace.

[0180] In this embodiment, the blast furnace charge distribution matrix is ​​a matrix that describes the material flow distribution and material surface changes during the blast furnace charge distribution process, and is usually constructed based on the information difference between real-time material surface information and standard material surface information.

[0181] In this embodiment, the initial blast furnace charge distribution model is a charge distribution data model selected from a charge distribution model database and matched with characteristic parameters of target blast furnace charge.

[0182] In this embodiment, the second material surface information is the material surface information output by the model after the blast furnace charge distribution matrix is ​​input into the initial blast furnace charge distribution model.

[0183] In this embodiment, the charge distribution optimization target refers to the optimization target expected to be achieved during the charge distribution process of the blast furnace, such as improving gas utilization rate, reducing coke ratio, etc.

[0184] In this embodiment, parameter adjustment refers to the process of adjusting the parameters in the blast furnace burden distribution matrix based on the comparison result.

[0185] In this embodiment, the first charge distribution model is established based on the first dimension reduction data and is a data model used to describe the charge distribution situation of the blast furnace.

[0186] The beneficial effect of the above technical solution is: by performing dimensionality reduction processing on the data generated in the blast furnace production process and conducting comprehensive analysis based on the data derivative results of the first distribution model, the gas utilization rate of the target blast furnace can be predicted more timely and accurately, so as to improve energy utilization and reduce enterprise energy consumption.

[0187] Embodiment 7:

[0188] Based on Example 5, a corresponding gas utilization model is established based on the blast furnace characteristics of the target blast furnace and the machine learning algorithm, and the gas utilization rate of the target blast furnace is predicted based on the first dimensionality reduction data and the first derivative data, including:

[0189] Step 41: Selecting a machine learning algorithm based on the blast furnace characteristics of the target blast furnace to construct a gas utilization model of the target blast furnace;

[0190] Step 42: inputting the first dimension reduction data and the first derivative data into the gas utilization rate model, thereby predicting the gas utilization rate of the target blast furnace based on each gas utilization rate model to obtain a first prediction result;

[0191] Step 43: Based on the first prediction result, data optimization is performed on corresponding data in the first production data set of the target blast furnace. At the same time, based on the first prediction result, matrix optimization is performed on the data matrix of the distribution matrix data in the distribution-related data of the target blast furnace, so as to improve the gas utilization rate of the target blast furnace.

[0192] In this embodiment, the gas utilization model is constructed based on a machine learning algorithm and is a model for predicting a target blast furnace gas utilization rate. The model can predict the gas utilization rate of a blast furnace based on input data (such as reduced dimensionality data, derived data, etc.).

[0193] In this embodiment, the first prediction result is a prediction result output by the model obtained after the first dimension reduction data and the first derivative data are input into the gas utilization rate model.

[0194] In this embodiment, data optimization refers to the process of adjusting and improving corresponding data in the first production data set of the target blast furnace based on the prediction results.

[0195] The beneficial effect of the above technical solution is that by processing process industrial data more accurately, the gas utilization rate of the target blast furnace can be predicted more timely and accurately to improve energy utilization.

[0196] Embodiment 8:

[0197] Based on Example 7, the gas utilization rate of the target blast furnace is improved, including:

[0198] Step 431: obtaining an expected target of real-time gas utilization rate of a target blast furnace, and comparing the expected target of real-time gas utilization rate with the first prediction result, thereby obtaining a first comparison result;

[0199] Step 432: Based on the first comparison result, identifying data whose impact on the gas utilization rate exceeds a preset minimum impact, as the first key data of the target blast furnace;

[0200] Step 433: extracting the comparison result corresponding to the first key data from the first comparison result to obtain a second comparison result;

[0201] Step 434: extracting a data optimization strategy corresponding to the second comparison result from the prediction-optimization database, and performing data optimization on the first key data, thereby improving the gas utilization rate of the target blast furnace;

[0202] Step 435: Based on the first comparison result and in combination with the first charging model of the target blast furnace, matrix optimization is performed on the data matrix of the charging matrix data in the charging-related data of the target blast furnace, thereby improving the gas utilization rate of the target blast furnace.

[0203] In this embodiment, the expected target of real-time gas utilization refers to the gas utilization level expected to be achieved during the operation of the blast furnace. This target is usually determined based on the operating conditions, operating experience and optimization requirements of the blast furnace.

[0204] In this embodiment, the first prediction result refers to a predicted value of a target blast furnace gas utilization rate obtained by using a gas utilization rate model.

[0205] In this embodiment, the first comparison result is a result obtained by comparing the expected target of real-time gas utilization rate with the first prediction result.

[0206] In this embodiment, the preset minimum impact level is a threshold of impact level set before data optimization. Only data whose impact level on gas utilization rate exceeds this threshold will be considered as key data for optimization.

[0207] In this embodiment, the first key data refers to data whose impact on gas utilization rate exceeds a preset minimum impact.

[0208] In this embodiment, the second comparison result is extracted from the first comparison result and is a comparison result corresponding to the first key data.

[0209] In this embodiment, the prediction-optimization database is a database storing a variety of prediction results, comparison results and corresponding data optimization strategies. This database is used to extract appropriate data optimization strategies based on the prediction results and comparison results.

[0210] In this embodiment, the data optimization strategy refers to a strategy for adjusting and improving key data in order to improve the utilization rate of blast furnace gas.

[0211] In this embodiment, data optimization refers to the process of adjusting and improving the first key data according to the data optimization strategy, aiming to improve the gas utilization rate of the target blast furnace.

[0212] The beneficial effect of the above technical solution is that by processing process industrial data more accurately, the gas utilization rate of the target blast furnace can be predicted more timely and accurately to improve energy utilization.

[0213] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. 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 gas utilization based on process industry data processing, characterized in that: include: Step 1: Collect various data of the target blast furnace, and establish real-time flow processing rules of the target blast furnace based on computer and database technology, so as to process and store various data of the target blast furnace, clean and complete the data, and obtain first processed data; Step 2: Based on the real-time unloading situation of the target blast furnace, the first processed data is decomposed and integrated, and the first production data set of the target blast furnace is obtained by combining the time lag effect of each data on the gas utilization rate in the decomposition and integration results; Step 3: extracting the monitoring data related to the target blast furnace body from the first production data set to obtain the first monitoring data, and performing data dimensionality reduction processing to obtain the first dimensionality reduction data, and at the same time obtaining the distribution-related data of the target blast furnace, thereby establishing the first distribution model of the target blast furnace, realizing data derivation, and obtaining the first derived data; Step 4: establishing a corresponding gas utilization model based on the blast furnace characteristics of the target blast furnace and the machine learning algorithm, and predicting the gas utilization of the target blast furnace based on the first dimensionality reduction data and the first derivative data; Wherein, step 2 includes: Step 21: Based on the real-time material unloading situation of the target blast furnace, the first processed data is decomposed, so that the charging data in the decomposed first processed data corresponds to the chemical composition and physical properties of the raw fuel data; Step 22: classify and integrate the remaining processed data of the target blast furnace except the charging data in the first processed data according to different data sources, so as to construct an initial production data set of the target blast furnace in combination with the processing result of the charging data; Step 23: Analyze whether each data in the initial production data set has a time lag effect on the gas utilization rate, determine the time lag time of the data with time lag effect based on a preset cross-correlation analysis method, and adjust the initial production data set according to the time lag time to obtain a first production data set; Wherein, step 3 includes: Step 31: extracting data related to the target blast furnace body from the first production data set to obtain first monitoring data; Step 32: performing data dimensionality reduction processing on the first monitoring data based on a preset dimensionality reduction method to obtain first dimensionality reduction data; Step 33: acquiring real-time charge distribution data of the target blast furnace, and extracting a suitable charge distribution data model from a charge distribution model database to construct a first charge distribution model in combination with charge distribution characteristics of the target blast furnace during charge distribution; Among them, real-time material distribution related data includes material batch feeding data and material distribution matrix data; Step 34: inputting the real-time charge distribution related data into the first charge distribution model, thereby obtaining the first derivative data of the target blast furnace; Wherein, step 4 includes: Step 41: Selecting a machine learning algorithm based on the blast furnace characteristics of the target blast furnace to construct a gas utilization model of the target blast furnace; Step 42: inputting the first dimension reduction data and the first derived data into the gas utilization rate model, thereby predicting the gas utilization rate of the target blast furnace based on each gas utilization rate model to obtain a first prediction result; Step 43: Based on the first prediction result, data optimization is performed on corresponding data in the first production data set of the target blast furnace. At the same time, based on the first prediction result, matrix optimization is performed on the data matrix of the distribution matrix data in the distribution-related data of the target blast furnace, so as to improve the gas utilization rate of the target blast furnace.

2. The method for predicting blast furnace gas utilization rate based on process industry data processing according to claim 1, characterized in that: Collect various data of the target blast furnace, and establish real-time flow processing rules of the target blast furnace based on computer and database technology, so as to process and store various data of the target blast furnace, clean and complete the data, and obtain the first processed data, including: Step 11: various data are collected in real time based on various sensors of the target blast furnace to obtain first collected data; Step 12: Determine a real-time data processing flow that matches the target blast furnace based on computer and database technology combined with the blast furnace characteristics of the target blast furnace, and determine a real-time data flow processing rule based on the data type of the blast furnace data and the first collected data; Step 13: Perform data conversion and data storage on the first collected data based on the real-time flow processing rule to obtain second collected data; Step 14: Clean and complete the second collected data to obtain the first processed data of the target blast furnace.

3. The method for predicting blast furnace gas utilization rate based on process industry data processing according to claim 1, characterized in that: Analyze whether each data in the initial production data set has a time lag effect on the gas utilization rate, determine the time lag time of the data with time lag effect based on a preset cross-correlation analysis method, and adjust the initial production data set according to the time lag time to obtain a first production data set, including: Step 231: obtaining the data type and data characteristics of each data in the initial production data set, thereby extracting the first historical data of the corresponding data type in the target blast furnace; Step 232: obtaining a first data reference based on the data characteristics of the first historical data and the data type related to the gas utilization rate, and selecting a corresponding first analysis method based on the first data reference; Step 233: determining whether there is a time lag effect on the gas utilization rate during the historical working process of the first historical data based on the first analysis method; If there is a time lag effect, the current data in the initial production data set is used as the first data; If there is no time lag effect, the current data in the initial production data set is used as the second data; Step 234: obtaining a first data set based on each first data integration of the initial production data set, and obtaining a second data set based on each second data integration of the initial production data set; Step 235: performing data processing on each first data in the first data set based on a preset cross-correlation analysis method, so as to determine a cross-correlation coefficient Ti corresponding to each first data when the lag time is i; ; Where Ti is the cross-correlation coefficient corresponding to the lag time i, is the observed value of the time series h at time ti, is the observed value of the time series s at time t, is the mean of the h time series, is the mean of the time series s, n is the number of observations in the time series, is the error affected by the sequence stability of the h time series, is the error affected by the sequence stability of the s time series, The error is caused by noise; Step 236: comparing the cross-correlation coefficients corresponding to each lag time, and extracting the lag time corresponding to the maximum cross-correlation coefficient as the first lag time of the current first data; Step 237: adjusting the relative time series of the current first data corresponding to the first data set based on the first time lag time, so as to achieve consistency of the first data in the time series, thereby obtaining a first adjusted data set; Step 238: Integrate the first adjustment data set with the second data set to obtain a first production data set of the target blast furnace.

4. The method for predicting blast furnace gas utilization rate based on process industry data processing according to claim 1, characterized in that: Acquire the real-time charge distribution related data of the target blast furnace, and extract the appropriate charge distribution data model from the charge distribution model database to construct the first charge distribution model in combination with the charge distribution characteristics of the target blast furnace during blast furnace charge distribution, including: Step 331: acquiring real-time charge distribution related data of the target blast furnace, and acquiring charge characteristic parameters of the target blast furnace for blast furnace charge distribution; Step 332: obtaining the information difference between the real-time material surface information and the standard material surface information, thereby constructing a blast furnace charging matrix of the target blast furnace; Step 333: selecting an initial blast furnace charge distribution model matching the target blast furnace from a charge distribution model database based on the charge characteristic parameters and charge information of the target blast furnace; Step 334: inputting the blast furnace charge distribution matrix into the initial blast furnace charge distribution model, thereby obtaining second charge surface information; Step 335: Obtain the charge distribution optimization target of the target blast furnace, thereby comparing the second charge surface information with the charge distribution optimization target, and adjusting the parameters of the blast furnace charge distribution matrix based on the comparison result, thereby verifying the initial blast furnace charge distribution model and obtaining a first charge distribution model.

5. The method for predicting blast furnace gas utilization rate based on process industry data processing according to claim 1, characterized in that: Based on the first prediction result, data corresponding to the first production data set of the target blast furnace is optimized. At the same time, based on the first prediction result, a matrix of the material distribution matrix data in the material distribution related data of the target blast furnace is optimized, thereby improving the gas utilization rate of the target blast furnace, including: Step 431: obtaining an expected target of real-time gas utilization rate of a target blast furnace, and comparing the expected target of real-time gas utilization rate with the first prediction result, thereby obtaining a first comparison result; Step 432: Based on the first comparison result, identifying data whose impact on the gas utilization rate exceeds a preset minimum impact, as the first key data of the target blast furnace; Step 433: extracting the comparison result corresponding to the first key data from the first comparison result to obtain a second comparison result; Step 434: extracting a data optimization strategy corresponding to the second comparison result from the prediction-optimization database, and performing data optimization on the first key data; Step 435: Based on the first comparison result and in combination with the first charging model of the target blast furnace, matrix optimization is performed on the data matrix of the charging matrix data in the charging-related data of the target blast furnace, thereby improving the gas utilization rate of the target blast furnace.

Images