Prediction method of blast furnace permeability index based on charge structure and metallurgical properties

Abstract

The present invention relates to a method for predicting the blast furnace permeability index based on charge structure and metallurgical properties. The method comprises the following steps: S1: obtaining chemical composition test values ​​of a number of sintered ore samples and corresponding droplet performance test data of the sintered ore; S2: constructing a prediction model A for the corresponding droplet performance index of the sintered ore based on the chemical composition of the sintered ore; S3: obtaining RDI+3.15 index data for low-temperature reduction pulverization of the sintered ore and constructing a prediction model B; S4: automatically obtaining actual values ​​through a detection device, iterating and updating prediction models A and B; S5: obtaining the RDI+3.15 index and corresponding droplet performance values ​​of pellets and lump ore; S6: obtaining real-time porosity; and S7: constructing a prediction model C based on comprehensive charge metallurgical properties, porosity, and coke particle size characteristic parameters. The blast furnace permeability index prediction model of the present invention has a prediction accuracy of over 92% within a range of ±5‰, providing a reference for actual blast furnace operation adjustments.

Description

Technical Field

[0001] The present invention relates to the technical field of metallurgy, and in particular to a method for predicting a blast furnace permeability index based on charge structure and metallurgical properties. Background Art

[0002] Blast furnace permeability refers to the ability of airflow to pass through the material layer under certain conditions. It directly determines whether the blast furnace gas flow distribution is reasonable, which ultimately affects whether the blast furnace production is stable and smooth. Therefore, in blast furnace production, the reasonable prediction and effective control of permeability, as one of the most core operational links, has always been a concern for ironmakers. Changes in blast furnace permeability are related to many factors in production, such as raw materials, air supply, and material distribution, which all have an impact on permeability. In recent years, with the improvement of blast furnace automation and management level, blast furnace operators have gradually paid attention to the problem of early prediction of permeability. However, there are still some problems in the current research on permeability prediction. For example, the established model only has offline calculation capabilities, or the selection of influencing factors only considers local and fixed parameters, which is difficult to adapt to industrial applications. Therefore, in actual production, operators still mostly rely on experience to control permeability, which has an adverse effect on the smooth operation of the blast furnace.

[0003] Currently, the iron-containing charge materials used in blast furnaces primarily consist of sintered ore, pelletized ore, and lump ore. Good metallurgical properties of iron-containing charge materials are a key factor influencing blast furnace permeability and an important prerequisite for achieving enhanced blast furnace smelting and energy conservation and emission reduction. Currently, domestic steel companies do not routinely test the droplet performance of comprehensive blast furnace charge materials. Most companies only test the droplet performance of a single type of iron-containing charge material, such as sintered ore, pelletized ore, and lump ore. Furthermore, metallurgical performance testing is costly and time-consuming. Consequently, ironmakers struggle to predict changing trends in blast furnace permeability based on changes in charge structure and metallurgical properties.

[0004] In recent years, the use of artificial intelligence technology to solve complex problems in blast furnaces has been widely used. Therefore, building a prediction and prejudgment system for the blast furnace permeability index based on the charge structure and metallurgical properties can provide a reference for optimizing the blast furnace charge structure and formulating a reasonable operating system in advance. It is also an important measure to achieve low-carbon ironmaking.

[0005] Patent (CN106777652A) discloses a method for predicting blast furnace permeability, including: collecting historical data from a blast furnace operation database; analyzing and preprocessing the historical data, selecting the historical data to obtain actual data that meets production requirements; obtaining the influencing factors of blast furnace permeability within the current time period and weighting the influencing factors according to their contribution; establishing a corresponding data set based on the weighted ranking of the permeability influencing factors; classifying the influencing factors in the data set according to the permeability parameters, and calculating the center of the influencing factors in each category; establishing several permeability prediction sub-models based on black box modeling; inputting adjusted raw material quality parameters or operating control parameters into the corresponding permeability prediction sub-models according to their corresponding category data as needed; and dynamically updating the model parameters using the latest data based on the frequency of data collection from the blast furnace operation database. This patent allows for dynamic updating of the permeability model, improving prediction accuracy.

[0006] The patent (CN 116124677A) discloses a method for rapidly evaluating the permeability of blast furnace sintered ore layers. The method includes the following steps: performing particle size analysis on n sintered ore samples; performing computer discrete element modeling on the analysis data of each group of sintered ore samples and performing uniformly distributed stacking; performing statistics on the stacked sintered ore layer to obtain the solid volume Vi1 / m3 and the stacking volume Vi0 / m3, and calculating the porosity Ei; performing computer modeling on the particle size composition of all sintered ore samples in the blast furnace in operation to obtain all the porosities E1, E2, E3, ..., En; performing multivariate linear regression analysis on the particle size composition and corresponding porosity of all the above sintered ore samples to obtain a porosity calculation formula, and calculating and predicting the porosity of new sintered ore samples. This method can quickly evaluate the permeability of sintered ore layers, provide early warning for blast furnace operation adjustments, and improve the scientificity, rationality, and accuracy of the permeability assessment and blast furnace operation adjustments.

[0007] However, there are no reports on patents regarding methods for predicting the blast furnace permeability index based on charge structure and metallurgical properties. Summary of the Invention

[0008] To solve the above problems, the present invention provides a blast furnace permeability index prediction method based on charge structure and metallurgical properties, which can provide a reference for optimizing blast furnace charge structure and formulating reasonable operating systems in advance, and is also an important measure to achieve low-carbon ironmaking.

[0009] The technical solution adopted by the present invention is: a blast furnace permeability index prediction method based on charge structure and metallurgical properties, characterized by comprising the following steps:

[0010] S1. Obtaining chemical composition test values ​​of a number of sintered ore samples and corresponding droplet performance test data of the sintered ore;

[0011] S2. Using the partial least squares multivariate analysis principle, a prediction model A for the corresponding sinter droplet performance index based on the main chemical composition of the sinter is constructed;

[0012] S3. Obtain the RDI+3.15 index data of the low-temperature reduction pulverization of sintered ore, and construct a prediction model B based on the correspondence between the ore blending structure parameters, the chemical composition of the sintered ore, and the sintering process control parameters and the RDI+3.15 index of the low-temperature reduction pulverization of sintered ore;

[0013] S4: The main chemical composition of the sintered ore before entering the furnace is automatically obtained through the LIBS online composition detection device and the sintered ore FeO online detection device. As new source data is generated, prediction models A and B automatically iterate and calculate and update the accumulated data to improve the accuracy of prediction models A and B.

[0014] S5. Obtain the RDI+3.15 index of pellets and lump ore and the corresponding droplet performance values ​​through phased tests;

[0015] S6. Obtaining the real-time porosity of the sintered ore, pelletized ore, and lump ore before entering the furnace;

[0016] S7. Construct a prediction model C for the blast furnace permeability index based on comprehensive charge metallurgical properties, porosity, and coke particle size characteristic parameters.

[0017] Preferably, in step S1, the test sample coverage range meets the following requirements: the mass fraction of SiO2 in the sintered ore is 4.53% to 6.09%; the mass fraction of FeO is 6.59% to 10.74%; the mass fraction of Al2O3 is 1.51% to 3.17%; the mass fraction of MgO is 1.02% to 2.65%; the R of the sintered ore is controlled between 1.72 and 2.35; the chemical composition of the sintered ore mainly comprises: TFe, SiO2, FeO, Al2O3, MgO and the basicity of the sintered ore R; the corresponding sintered ore droplet properties mainly comprise: softening start temperature T10, softening end temperature T40, melting start temperature Ts, dripping temperature Td, softening range ΔT1=T40-T10, and melting range ΔT1=Td-Ts.

[0018] Preferably, the specific steps of step S2 are: using the partial least squares multivariate analysis principle, taking the main component parameters of the sintered ore: SiO2, FeO, Al2O3, MgO and R as independent variables, and the detection data of various indicators of the high-temperature soft melting dripping performance of the sintered ore as dependent variables for multivariate regression, and constructing a prediction model A based on the corresponding sintered ore droplet performance indicators based on the main chemical components of the sintered ore.

[0019] Preferably, the specific steps of step S3 are: obtaining the RDI+3.15 index data of low-temperature reduction and pulverization of sintered ore, the sintering corresponding ore blending structure, the main components of the sintered ore and the sintering process control parameter data, taking the ore blending structure parameters, the chemical composition of the sintered ore, and the sintering process control parameters as independent variables, and the sintered ore RDI+3.15 detection data as the dependent variable for multiple regression, and constructing a prediction model B based on the correspondence of the ore blending structure parameters, the chemical composition of the sintered ore and the sintering process control parameters to the RDI+3.15 index of low-temperature reduction and pulverization of sintered ore.

[0020] Preferably, the ore distribution structure parameters are the ore distribution structure ratios of each pile of mixed ore, and the sintering process control parameters include exhaust gas temperature, ignition temperature, return ore ratio and sintering process negative pressure.

[0021] Preferably, in step S4, the relative error between the predicted result and the actual value of the start softening temperature T10 is 1.05%; the relative error between the predicted result and the actual value of the softening end temperature T40 is 0.66%; the relative error between the predicted result and the actual value of the dripping temperature Td is 0.84%; the relative error between the predicted result and the actual value of the start melting temperature Ts is 0.32%; the relative error between the predicted result and the actual value of the sintered ore soft melting range is 6.70%; the relative error between the predicted result and the actual value of the sintered ore melting range is 6.79%; the relative error between the predicted result and the actual value of the sintered ore RDI+3.15 is 3.60%.

[0022] Preferably, in step S6, industrial CT is used to obtain the real-time porosity of the sintered ore, pelletized ore, and lump ore before entering the furnace, the scanned slice data is processed by three-dimensional reconstruction software, and the porosity of the sintered ore, pelletized ore, and lump ore is extracted by threshold segmentation.

[0023] Preferably, the specific steps of step S7 are: constructing a prediction model C for the blast furnace permeability index based on the comprehensive metallurgical properties, porosity and coke particle size characteristic parameters. In order to avoid the generalization problem of the prediction model C caused by the introduction of too many independent variables, several comprehensive independent variables are set to further construct a regression prediction model C between the blast furnace permeability index and the comprehensive charge RDI index, comprehensive charge T10 index, comprehensive charge T40 index, comprehensive charge Td index, comprehensive charge Ts index, comprehensive charge ΔT1 index, comprehensive charge ΔT2 index, comprehensive charge porosity ε index and coke particle size index, so as to guide the blast furnace operation adjustment and the optimization adjustment control of the front-end sintering metallurgical performance.

[0024] The beneficial effects achieved by the present invention are:

[0025] 1. Currently, the iron-containing charge materials used in blast furnaces primarily include sintered ore, pellets, and lump ore. Generally speaking, the types of pellets and lump ore used in ironmaking plants' blast furnaces are relatively simple, and their chemical compositions are relatively stable. Therefore, through periodic testing, the RDI+3.15 index and corresponding droplet properties (softening start temperature, softening end temperature, melting start temperature, dripping temperature, softening range, and melting range) of pellets and lump ore can be obtained.

[0026] 2. The main chemical composition of the sintered ore before entering the furnace can be automatically obtained in real time using the LIBS online composition detection device and the sintered ore FeO online detection device. As new source data is generated, the model can automatically iteratively run and update the above prediction formula, thus further improving the model's prediction accuracy as the amount of data accumulates. In the sample space, the relative error between the predicted and actual values ​​of the softening start temperature T10 is 1.05%; the relative error between the predicted and actual values ​​of the softening end temperature T40 is 0.66%; the relative error between the predicted and actual values ​​of the dripping temperature Td is 0.84%; the relative error between the predicted and actual values ​​of the melting start temperature Ts is 0.32%; the relative error between the predicted and actual values ​​of the sintered ore softening range is 6.70%; the relative error between the predicted and actual values ​​of the sintered ore melting range is 6.79%; and the relative error between the predicted and actual values ​​of the sintered ore RDI+3.15 is 3.60%. The error of the model prediction result is small, and the model can be used to predict actual production;

[0027] 3. The blast furnace permeability index prediction model of the present invention is based on the charge structure and metallurgical properties. Its prediction accuracy of ±5‰ reaches more than 92%, so this model can provide a reference basis for the actual operation adjustment of the blast furnace. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the present invention;

[0029] Figure 2-8 This is a comparison chart of the predicted values ​​and actual values ​​of blast furnace permeability prediction model A and prediction model B. DETAILED DESCRIPTION

[0030] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0031] like Figure 1As shown, a method for predicting the blast furnace permeability index based on charge structure and metallurgical properties of the present invention comprises the following steps:

[0032] S1. Obtain chemical composition test values ​​for a number of sinter ore samples and corresponding droplet performance test data. The test samples must meet the following requirements: the mass fraction of SiO2 in the sinter ore is 4.53% to 6.09%; the mass fraction of FeO is 6.59% to 10.74%; the mass fraction of Al2O3 is 1.51% to 3.17%; the mass fraction of MgO is 1.02% to 2.65%; and the R of the sinter ore is controlled between 1.72 and 2.35. The chemical composition of the sinter ore mainly includes: TFe, SiO2, FeO, Al2O3, MgO, and the basicity R of the sinter ore. The corresponding sinter droplet properties are mainly: softening start temperature (T10), softening end temperature (T40), softening start temperature (Ts), dripping temperature (Td), softening range (ΔT1 = T40-T10), melting range (ΔT1 = Td-Ts); the specific test sample sinter ore chemical composition test values ​​are as follows

[0033] As shown in Table 1:

[0034]

[0035]

[0036] Table 1 Chemical composition of the experimental sintered ore samples

[0037] By testing the high-temperature metallurgical properties of the above-mentioned sample sintered ore, the corresponding sintered ore softening start temperature (T10), softening end temperature (T40), melting start temperature (Ts), dripping temperature (Td), softening range (ΔT1=T40-T10), and melting range (ΔT1=Td-Ts) test results were obtained, as shown in Table 2:

[0038]

[0039]

[0040] Table 2 Droplet performance parameters of different sintered ore samples

[0041] S2. Using the principle of partial least squares multivariate analysis, the main component parameters of sintered ore (SiO2, FeO, Al2O3, MgO and R) were used as independent variables, and the test data of various indicators of sintered ore high-temperature soft melting dripping performance were used as dependent variables for multivariate regression. A prediction model A corresponding to the sintered ore droplet performance indicators based on the main chemical components of the sintered ore was constructed. The obtained prediction model A is shown in Table 3:

[0042]

[0043] Table 3 Regression formula of sintered ore soft melting dripping performance and variable projection importance index value

[0044] S3. Obtain data on the RDI+3.15 index of low-temperature reduction and pulverization of sintered ore, the ore blending structure corresponding to sintering, the main components of the sintered ore, and sintering process parameters. Use the ore blending parameters (mainly the ore blending structure ratio of each mixed ore pile), the chemical composition of the sintered ore, and the sintering process control parameters (mainly including exhaust gas temperature, ignition temperature, return ore ratio, and sintering process negative pressure) as independent variables, and the sintered ore RDI+3.15 test data as the dependent variable for multiple regression. Construct a prediction model B based on the correspondence between the ore blending parameters, the chemical composition of the sintered ore, and the sintering process control parameters and the RDI+3.15 index of low-temperature reduction and pulverization of sintered ore. The obtained prediction model B is detailed in Table 4:

[0045]

[0046]

[0047] Table 4 Independent variables used in the sinter RDI prediction model

[0048] The prediction model formula of the RDI+3.15 index of sintered ore low-temperature reduction pulverization is as follows:

[0049] Sintered ore RDI +3.15 =99.64-0.296X1-1.06X2-11.57X3-0.34X4-31.86X5+0.74X6-0.004X7-0.03X8-0.007X9-0.014X 10 +0.19X 11 +0.09X 12 +0.12X 13 -0.04X 14 -0.004X 15 +0.01X 16 -0.053X 17 -0.046X 18 +0.007X 19 +0.003X 20 -0.046X 21 -0.086X 22 -0.142X 23 -0.195X 24 -0.152X 25 +0.165X 26 +0.336X 27 +0.033X 28 +0.023X29 +0.705X 30 +1.07X 31 +0.127X 32

[0050] Where: The parameters represented by the variables X1 to X32 are shown in Table 4;

[0051] S4. The main chemical composition of the sintered ore before entering the furnace can be automatically obtained in real time through the LIBS online component detection device and the sintered ore FeO online detection device. As new source data is generated, the prediction model A and the prediction model B can be automatically iterated and updated. Therefore, as the amount of data accumulates, the accuracy of the prediction model A and the prediction model B can be further improved. Figure 2-8 As shown in the figure: (a) is T 10 ; (b) is T 40 ; (c) is T d ; (d) is T s ; (e) is the soft melting range; (f) is the melting range; (g) is RDI +3.15 1 is the actual value; 2 is the predicted value. In the above sample space, the relative error between the predicted result and the actual value of the softening start temperature T10 is 1.05%; the relative error between the predicted result and the actual value of the softening end temperature T40 is 0.66%; the relative error between the predicted result and the actual value of the dripping temperature Td is 0.84%; the relative error between the predicted result and the actual value of the melting start temperature Ts is 0.32%; the relative error between the predicted result and the actual value of the sinter ore softening range is 6.70%; the relative error between the predicted result and the actual value of the sinter ore melting range is 6.79%; the relative error between the predicted result and the actual value of the sinter ore RDI+3.15 is 3.60%. The model prediction result has a small error, and the model can be used to predict actual production.

[0052] S5. Obtain the RDI+3.15 index of pellets and lump ores and the corresponding droplet performance values ​​through phased tests. Currently, the iron-containing charges for blast furnaces mainly include sintered ore, pellets, and lump ores. Generally speaking, the types of pellets and lump ores used in blast furnaces of ironmaking plants are relatively simple and their chemical compositions are relatively stable. Therefore, the RDI+3.15 index of pellets and lump ores and the corresponding droplet performance values ​​(softening start temperature, softening end temperature, softening start temperature, dripping temperature, softening range, melting range) can be obtained through phased tests.

[0053] S6. Obtain the real-time porosity of the sintered ore, pellets, and lump ore before entering the furnace; obtain the real-time porosity of the sintered ore, pellets, and lump ore before entering the furnace using industrial CT. The industrial CT is installed on the blast furnace charging belt. X-ray computed tomography (CT) is a relatively mature three-dimensional scanning technology. The scanned slice data is processed using three-dimensional reconstruction software, and the porosity of the sintered ore, pellets, and lump ore is extracted through threshold segmentation.

[0054] S7. Construct a prediction model C for the blast furnace permeability index based on the comprehensive metallurgical properties, porosity and coke particle size characteristic parameters; construct a prediction model C for the blast furnace permeability index based on the comprehensive metallurgical properties, porosity and coke particle size characteristic parameters. In order to avoid the generalization problem of the prediction model C caused by the introduction of too many independent variables, several comprehensive independent variables are set to further construct a regression prediction model C between the blast furnace permeability index and the comprehensive charge RDI index, comprehensive charge T10 index, comprehensive charge T40 index, comprehensive charge Td index, comprehensive charge Ts index, comprehensive charge ΔT1 index, comprehensive charge ΔT2 index, comprehensive charge porosity ε index and coke particle size index, so as to achieve guidance on blast furnace operation adjustment and optimization adjustment control of front-end sintering metallurgical performance;

[0055] The independent variables selected in the modeling process are shown in Table 5:

[0056]

[0057]

[0058] Table 5 Independent variables selected for modeling

[0059] Due to the large number of independent variables, the regression process is more complicated. Therefore, several new independent variables Z (comprehensive independent variables) are introduced in the modeling process, and their calculations are shown in the following formulas. After the dependent variable Y is regressed with several comprehensive independent variables Z, the original complex multiple regression problem is transformed into a simpler regression problem, avoiding the model generalization problem that may be caused by the introduction of too many independent variables.

[0060] The above-mentioned comprehensive independent variables Z are defined as follows:

[0061] The comprehensive charge RDI index corresponds to the independent variable Z1:

[0062]

[0063] The comprehensive charge T10 index corresponds to the independent variable Z2:

[0064]

[0065] Comprehensive charge T40 corresponds to independent variable Z3:

[0066]

[0067] The comprehensive charge Td index corresponds to the independent variable Z4:

[0068]

[0069] The comprehensive charge Ts index corresponds to the independent variable Z5:

[0070]

[0071] The comprehensive charge ΔT1 index corresponds to the independent variable Z6:

[0072]

[0073] The comprehensive charge ΔT2 index corresponds to the independent variable Z7:

[0074]

[0075] The comprehensive charge porosity ε index corresponds to the independent variable Z8:

[0076]

[0077] Coke particle size corresponds to the independent variable Z9:

[0078] Z9=x 28 ·x 29

[0079] The blast furnace permeability index is taken as the dependent variable Y, and the above new independent variables Z1 to Z9 are used to establish a regression prediction model C; the obtained blast furnace permeability index regression model is as follows:

[0080]

[0081]

[0082] The above model predicts a hit rate of ±5‰ of more than 92%, so the prediction model C can provide a reference for the actual operation adjustment of the blast furnace.

[0083] It should be noted that the description of the above technical solutions is illustrative only. This specification may be embodied in various forms and should not be construed as limiting the technical solutions set forth herein. Rather, these descriptions are provided to ensure that the disclosure of the present invention is thorough and complete and to fully convey the scope of the disclosure to those skilled in the art. Furthermore, the technical solutions of the present invention are limited only by the scope of the claims.

[0084] Finally, it should be noted that the above embodiments are merely representative examples of the present invention. Obviously, the present invention is not limited to the above embodiments and is susceptible to numerous variations. Any simple modifications, equivalent variations, and modifications to the above embodiments based on the technical essence of the present invention shall be deemed to fall within the scope of protection of the present invention.

Claims

1. A method for predicting blast furnace permeability index based on charge structure and metallurgical properties, characterized by: The following steps are involved: S1. Obtaining chemical composition test values ​​of a number of sintered ore samples and corresponding droplet performance test data of the sintered ore; S2. Using the partial least squares multivariate analysis principle, a prediction model A for the corresponding sinter droplet performance index based on the main chemical composition of the sinter is constructed; S3. Obtain the RDI+3.15 index data of the low-temperature reduction pulverization of sintered ore, and construct a prediction model B based on the correspondence between the ore blending structure parameters, the chemical composition of the sintered ore, and the sintering process control parameters and the RDI+3.15 index of the low-temperature reduction pulverization of sintered ore; S4: The chemical composition of the sintered ore before entering the furnace is automatically obtained through the LIBS online composition detection device and the sintered ore FeO online detection device. As new source data is generated, prediction models A and B automatically iterate and calculate and update the accumulated data to improve the accuracy of prediction models A and B. S5. Obtain the RDI+3.15 index of pellets and lump ore and the corresponding droplet performance values ​​through phased tests; S6. Obtaining the real-time porosity of the sintered ore, pelletized ore, and lump ore before entering the furnace; S7. Construct a prediction model C for the blast furnace permeability index based on comprehensive charge metallurgical properties, porosity and coke particle size characteristic parameters; in order to avoid the generalization problem of the prediction model C caused by the introduction of too many independent variables, set a number of comprehensive independent variables, and further construct a regression prediction model C between the blast furnace permeability index and the comprehensive charge RDI index, comprehensive charge T10 index, comprehensive charge T40 index, comprehensive charge Td index, comprehensive charge Ts index, comprehensive charge ΔT1 index, comprehensive charge ΔT2 index, comprehensive charge porosity ε index and coke particle size index, so as to guide the blast furnace operation adjustment and the optimization adjustment and control of the front-end sintering metallurgical properties; wherein the comprehensive charge RDI index is composed of the RDI+3.15 index of sintered ore, pelletized ore and lump ore, the sintered ore RDI+3.15 index is obtained by step S3, and the pelletized ore and lump ore RDI+3.15 index are obtained by step S5.

2. The method for predicting blast furnace permeability index based on charge structure and metallurgical properties according to claim 1, characterized in that: In step S1, the test sample coverage range meets the following requirements: the mass fraction of SiO2 in the sintered ore is 4.53% to 6.09%; the mass fraction of FeO is 6.59% to 10.74%; the mass fraction of Al2O3 is 1.51% to 3.17%; the mass fraction of MgO is 1.02% to 2.65%; the R of the sintered ore is controlled between 1.72 and 2.35; the chemical composition of the sintered ore mainly includes: TFe, SiO2, FeO, Al2O3, MgO and the basicity of the sintered ore R; the corresponding sintered ore droplet properties are mainly: softening start temperature T10, softening end temperature T40, melting start temperature Ts, dripping temperature Td, softening range ΔT1=T40-T10, and melting range ΔT1=Td-Ts.

3. The method for predicting blast furnace permeability index based on charge structure and metallurgical properties according to claim 1, characterized in that: The specific steps of step S2 are: using the partial least squares multivariate analysis principle, taking the main component parameters of sintered ore: SiO2, FeO, Al2O3, MgO and R as independent variables, and the detection data of various indicators of sintered ore high-temperature soft melting dripping performance as dependent variables for multivariate regression, and constructing a prediction model A based on the corresponding sintered ore droplet performance indicators based on the main chemical components of the sintered ore.

4. The method for predicting blast furnace permeability index based on charge structure and metallurgical properties according to claim 1, wherein: The specific steps of step S3 are: obtaining the RDI+3.15 index data of low-temperature reduction and pulverization of sintered ore, the ore blending structure corresponding to sintering, the main components of the sintered ore, and the sintering process control parameter data, taking the ore blending structure parameters, the chemical composition of the sintered ore, and the sintering process control parameters as independent variables, and the sintered ore RDI+3.15 detection data as the dependent variable for multiple regression, and constructing a prediction model B based on the correspondence between the ore blending structure parameters, the chemical composition of the sintered ore, and the sintering process control parameters and the RDI+3.15 index of low-temperature reduction and pulverization of sintered ore.

5. The method for predicting blast furnace permeability index based on charge structure and metallurgical properties according to claim 4, characterized in that: The ore blending structure parameters are the ore blending structure ratio of each pile of mixed ore, and the sintering process control parameters include exhaust gas temperature, ignition temperature, return ore ratio and sintering process negative pressure.

6. The method for predicting blast furnace permeability index based on charge structure and metallurgical properties according to claim 1, characterized in that: In step S4, the relative error between the predicted result and the actual value of the softening start temperature T10 is 1.05%; the relative error between the predicted result and the actual value of the softening end temperature T40 is 0.66%; the relative error between the predicted result and the actual value of the dripping temperature Td is 0.84%; the relative error between the predicted result and the actual value of the melting start temperature Ts is 0.32%; the relative error between the predicted result and the actual value of the sintered ore softening range is 6.70%; and the relative error between the predicted result and the actual value of the sintered ore melting range is 6.79%. The relative error between the predicted result of sintered ore RDI+3.15 and the actual value is 3.60%.

7. The method for predicting blast furnace permeability index based on charge structure and metallurgical properties according to claim 1, characterized in that: In step S6, industrial CT is used to obtain the real-time porosity of the sintered ore, pelletized ore, and lump ore before entering the furnace. The scanned slice data is processed by 3D reconstruction software, and the porosity of the sintered ore, pelletized ore, and lump ore is extracted by threshold segmentation.

Images