Pellet metallurgy performance prediction method based on colorimetry principle

Abstract

The invention provides a pellet metallurgical performance prediction method based on a colorimetric principle, and relates to the technical field of pellet metallurgical performance detection. The method comprises the steps of pellet surface pretreatment, pellet pre-classification, pellet colorimetry data acquisition, pellet metallurgical performance determination, linear regression model establishment, performance inspection, data set transformation and new linear regression model establishment. According to the method, the colorimetric parameters are collected, the metallurgical performance of the pellets is measured, the linear regression equation model is established, and the metallurgical performance of the pellets is matched and predicted according to the linear regression equation model, so that the defects of danger, time consumption and the like during traditional metallurgical performance measurement are overcome, and the method has important significance in promoting pellet detection and analysis means progress, safe production and efficient intelligent production; the method is simple and easy to operate, green and environment-friendly, low in cost, short in process, high in efficiency and beneficial to industrial large-scale production and popularization.

Description

Technical Field

[0001] The present invention relates to the technical field of detecting the metallurgical properties of pellet ore, and particularly to a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry. Background Art

[0002] In order to save energy and reduce emissions at the source, iron and steel enterprises continuously increase the proportion of pellet ore charged into the blast furnace during the blast furnace ironmaking process to assist in achieving the goal of large-scale pellet smelting in the blast furnace. High-quality finished pellet ore requires reasonable roasting parameters, excellent raw material components, and advanced equipment to meet the ironmaking policy of high efficiency, high quality, low consumption, long life, and environmental protection, and to help iron and steel enterprises improve production efficiency, reduce energy consumption and emissions, and achieve sustainable development.

[0003] At present, the detection and evaluation of the metallurgical properties of finished pellet ore mainly rely on high-temperature experimental detection in a simulated blast furnace environment. The experimental steps are relatively complex and time-consuming, and there is a lack of quantitative indicators that can effectively characterize the metallurgical properties of pellet ore only through simple measurement and general experience.

[0004] For example: The master's degree thesis of North China University of Science and Technology - Research on the Improved SVM Model for Predicting the Metallurgical Properties of Pellet Ore discloses a model prediction method for the properties of pellet ore. The research is carried out based on the fact that "the microstructure of pellet ore determines its metallurgical properties, and the metallurgical properties reflect its microstructure". The improved SVM algorithm and model are obtained by organically coupling the adaptive selection of SVM kernel parameters and the adaptive combination of SVM kernel types, and are compared with 9 control algorithms; obviously, the acquisition method of this model is complex and requires obtaining the microstructure of pellet ore first, including electron microscope images of various parts of pellet ore, ore phase texture characteristics, ore phase color characteristics, ore phase fractal characteristics, etc. Although the prediction is accurate, the prediction cost is high and the efficiency is low.

[0005] The master's degree thesis of Northeastern University - Research and Application of the Quality Prediction Model of Pellet Ore Based on Data discloses a system capable of predicting the quality of grate-kiln pellet ore. The research is carried out on three types, namely BP, GA-BP, and PSO-BP, and a prediction model is established. After being verified by MATLAB simulation, the system is finally obtained based on the VS2008 platform, MATLAB, and Microsoft Access; obviously, this system focuses on the relationship between various process parameters and the quality index of finished pellet ore and does not provide technical inspiration for predicting the metallurgical properties of pellet ore.

[0006] Chinese Patent CN113793308A discloses an intelligent rating method and device for the quality of pellet ore based on a neural network. The method first obtains a light microscope pellet ore picture and performs median filtering on the light microscope pellet ore picture; then performs HSV transformation and Gamma correction on the median-filtered light microscope pellet ore picture; then inputs the corrected light microscope pellet ore picture into an improved U-Net neural network to obtain the microscopic features of the light microscope pellet ore picture; finally, inputs the microscopic features into a BP neural network to obtain the corresponding grade of the pellet ore; obviously, it is necessary to initially establish the relationship between the microscopic morphology and macroscopic properties of pellet ore through neural network and image processing technology on the basis of pellet ore; and all these show the complexity of the process, the increased operation difficulty, and the long time consumption. Summary of the Invention

[0007] In order to solve the technical problems in the evaluation of the metallurgical properties of pellet ore in the prior art, such as high prediction cost, low efficiency, many influencing factors in the prediction process, great operation difficulty in the prediction process, and the prediction model of intelligent algorithms based on the ore phase texture, ore phase composition, and electron microscope pictures of pellet ore that require a large amount of measurement data; the present invention proposes a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry that can solve the foregoing problems. The technical solution is as follows:

[0008] A method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry, and the method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry includes the following steps:

[0009] S1. Surface pretreatment of pellet ore: Perform surface pretreatment on the selected pellet ore to obtain pellet ore with a clean surface;

[0010] S2. Pre-classification of pellet ore: Perform manual pre-classification and marking on the pellet ore with a clean surface in S1 to obtain pre-classified pellet ore;

[0011] S3. Data collection of pellet ore colorimetry: Use a colorimeter to repeatedly collect the colorimetry data L*, a*, b* of pellet ore with different colors in S2 according to the same recognition standard for multiple times, and take the respective averages to establish a data set corresponding to different pellet colors;

[0012] Among them, L* is the brightness value, a* is the red-green value, and b* is the yellow-blue value;

[0013] S4. Determination of the metallurgical properties of pellet ore: Perform multiple metallurgical property determinations on the pellet ore corresponding to the data set of different pellet colors in S3 after non-destructive testing, take the average value as the required value and record it to obtain a performance data set corresponding to the data set of different pellet colors;

[0014] S5. Establishment of linear regression model: Conduct collinearity test on the datasets of different pellet colors in S4, select appropriate colorimetric indicators as independent variables, match the selected colorimetric data with the metallurgical property results of the pellets, and establish a linear regression model.

[0015] S6. Performance test: Test the R of the linear regression model in S5 2 and the significance level, and predict and identify the metallurgical properties of the new pellets based on this linear regression model.

[0016] S7. Transformation of the dataset and establishment of a new linear regression model: After changing the ore blending scheme and process regime, repeat the above S1 - S6 process and establish a linear regression model for the new process.

[0017] Optionally, the surface pretreatment method in S1 is as follows:

[0018] (1) Wash the powder on the surface of each pellet with distilled water to remove the impurities and dust adhering to the ore surface.

[0019] (2) Put the washed pellets into a drying oven to dry, thereby removing the surface moisture.

[0020] Optionally, the drying in S1 is carried out in a drying oven at 105 °C for 3 h.

[0021] Optionally, removing the impurities and dust adhering to the ore surface in S1 mainly refers to removing the red hematite ore powder adhering to the surface of the blue pellets.

[0022] Optionally, the manual pre - classification and marking in S2 are mainly carried out by using Arabic numerals or letters or a combination of letters and numbers to distinguish and mark according to the different colors on the surface of the pellets, so as to facilitate the chromaticity - property matching and type distinction in the subsequent experimental process.

[0023] Optionally, when the L* value in S3 is 0, it represents black, and when it is 100, it represents white; the a* value represents the degree of red - green. If the a* value is positive, it means the color is more red, and if the a* value is negative, it means the color is more green; the b* value represents the degree of yellow - blue. If the b* value is positive, it means the color is more yellow, and if the b* value is negative, it means the color is more blue.

[0024] Optionally, the measurement methods of the colorimetric data L*, a*, b* of the pellets of different colors in S3 are as follows:

[0025] (1) Start the colorimeter, use the standard background tool, and calibrate the colorimeter twice under white and black backgrounds.

[0026] (2) Place the pellet in the colorimetric measurement area of the colorimeter. The colorimeter emits light by itself. After the light shines on the surface of the pellet, it is reflected, and thus the target parameters L*, a*, b* are obtained.

[0027] (3) To comprehensively consider the overall chromaticity index and performance of pellet ore and avoid contingency, for each pellet ore, randomly rotate it under the chromaticity meter and measure the chromaticity at five positions twice each. Take the average of the ten obtained chromaticity values as the final colorimetric index of the pellet.

[0028] (4) Due to the non-destructive nature of the chromaticity meter measurement, each measured pellet can be individually numbered and stored classified.

[0029] Optionally, the chromaticity meter parameters measured in S3 are uniformly set to the SCI measurement mode, apply the D65 light source, and select a 10° field of view for observation.

[0030] Optionally, the properties measured in the metallurgical property determination in S4 include pellet compressive strength, reducibility, and reduction swelling.

[0031] Optionally, for the establishment of the linear regression model in S5, it is necessary to select one or more appropriate independent variables from the previously measured colorimetric data of the chromaticity indices L*, a*, and b*, and match them with the results of the metallurgical properties of the pellet ore.

[0032] Optionally, there is a collinearity inflation factor VIF detected among L*, a*, and b* in S5, which proves that the collinearity degree of the dependent variables is serious. A linear regression equation cannot be established with these values, and other optimal multiple independent variables need to be selected.

[0033] Optionally, the performance test in S6 is specifically as follows: taking the chromaticity index of the pellet ore as the independent variable and the metallurgical property as the dependent variable, establish a linear regression equation, and evaluate the obtained R 2 and the significance level. R 2 represents the goodness of fit, and the significance level represents the probability of a small-probability event occurring in one experiment. It is necessary to evaluate whether it meets the requirements according to the actual situation and production requirements, so as to predict the metallurgical properties of the pellet ore.

[0034] Optionally, R 2 The closer the value is to the value 1, the better the fitting effect.

[0035] Optionally, in the present invention, the purpose is to predict the metallurgical properties of pellet ore according to different colors. In fact, the color of the pellet ore is distributed between blue and red. However, due to the mutual friction between the pellet ores, powder dropping will occur. It is possible that the powder on other red pellet ores will adhere to other blue pellet ores, affecting their original color, and then affecting the chromaticity value of the colorimetric measurement result, and further causing errors in the experimental results. Therefore, it is necessary to perform pretreatment and cleaning first to remove the red powder on the blue pellet ores and make them return to their intrinsic colors.

[0036] The above technical solution has at least the following beneficial effects compared with the prior art:

[0037] In the above solution, the present invention provides a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry, which can solve the technical problems existing in the evaluation of the metallurgical properties of pellet ore, such as high prediction cost, low efficiency, many influencing factors in the prediction process, great operation difficulty in the prediction process, and prediction models of intelligent algorithms based on the ore phase patterns, ore phase compositions, and SEM images of pellet ore that require a large amount of measurement data.

[0038] Through surface pretreatment, the present invention can avoid different-color pellet ores from being extruded, rubbed, and powdered against each other during transportation and covering the surfaces of other pellet ores, which affects the observation and measurement of their surface colors. Therefore, during the test, the intrinsic color data of pellet ore is obtained through surface pretreatment to eliminate the interference of mutual friction and powdering on color observation and measurement.

[0039] Through the pre-classification of pellet ore, the present invention can pre-distinguish and label different-color pellet ores after preliminary treatment, which is convenient for type distinction and chromaticity-performance matching in subsequent experimental processes.

[0040] Through the data acquisition of the colorimetry of pellet ore, the present invention can quantitatively characterize the color of pellet ore through three colorimetry indexes of L*, a*, and b*, finally establish the relationship between them and the metallurgical properties, and test their correlation to obtain a linear regression prediction model.

[0041] Through the determination of the metallurgical properties of pellet ore, the present invention can match and correspond the colorimetry indexes of different-color pellet ores with the metallurgical properties (such as compressive strength, reducibility, and reduction swelling performance). According to the metallurgical property difference system under different chromaticity indexes, a performance evaluation prediction model based on colorimetry is built, so as to solve the disadvantages of complex and time-consuming on-site detection, and promote safe production and high-efficiency intelligent production.

[0042] Through the establishment of a linear regression model, the present invention can match and construct the correlation between the colorimetry indexes of finished pellet ore and its various metallurgical properties, build a metallurgical property evaluation system of pellet ore based on colorimetry, and then predict its metallurgical properties through chromaticity indexes and regression equations, which helps iron and steel enterprises improve production efficiency in the performance detection of pellet ore.

[0043] Through performance inspection, the present invention can compare the true metallurgical properties of finished pellet ore with the metallurgical properties of finished pellet ore predicted based on the principle of colorimetry, test the deviation degree, and evaluate whether it meets the requirements according to the actual situation and production requirements, so as to predict the metallurgical properties of pellet ore.

[0044] The linear regression model prediction and identification obtained by designing the present invention, which conforms to the relationship between the metallurgical properties of pellet ore and Lab colorimetry values, has shown significant innovation and improvement in the field of pellet ore detection and analysis, and has a positive effect on cost reduction, quality improvement and efficiency increase of iron and steel enterprises, and promoting the optimization of detection technology.

[0045] The method of the present invention not only solves the disadvantages such as danger and time-consuming in traditional metallurgical property measurement, but also is of great significance for promoting the progress of pellet ore detection and analysis means, safe production and high-efficiency intelligent production; in addition, it can provide technical support for the optimization of the process system in the pellet ore production process, quality improvement and the further development of the prospect goal of large-scale pellet ore smelting in blast furnaces.

[0046] In summary, compared with other traditional methods, the method of the present invention collects colorimetry parameters and measures the metallurgical properties of pellet ore, establishes a linear regression equation model and predicts its metallurgical properties accordingly, solves the disadvantages such as danger and time-consuming in traditional metallurgical property measurement, and is of great significance for promoting the progress of pellet ore detection and analysis means, safe production and high-efficiency intelligent production; the method is simple and easy to operate, environmentally friendly, low-cost, short-process and high-efficiency, and is conducive to large-scale industrial production and promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0048] Figure 1 is the technical roadmap of a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry of the present invention;

[0049] Figure 2 is the physical diagram of two different color pellet ores of a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry of the present invention, and the differences between different color pellet ores are large; among them, the first row is Example 1, a is a relatively red pellet ore, and b is a relatively blue pellet ore; the second row is Example 2, c is a relatively red pellet ore, and d is a relatively blue pellet ore;

[0050] Figure 3 is the linear regression difference analysis diagram of a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry in Example 1 of the present invention;

[0051] Figure 4 is the linear regression difference analysis diagram of a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry in Example 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] The technical solutions in the present invention will be described below with reference to the accompanying drawings.

[0053] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of the word "example" is intended to present concepts in a specific manner. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two.

[0054] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference is not emphasized, the meanings they express are the same.

[0055] In the embodiments of the present invention, sometimes a subscript such as W 1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meanings they express are the same.

[0056] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0057] A method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry, the method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry combines Figure 1 the following steps:

[0058] S1. Surface pretreatment of pellet ore: Perform surface pretreatment on the selected pellet ore to obtain pellet ore with a clean surface;

[0059] S2. Preliminary classification of pellet ore: Perform manual preliminary classification and marking on the pellet ore with a clean surface in S1 to obtain preliminarily classified pellet ore;

[0060] S3. Data acquisition of pellet ore colorimetry: Use a colorimeter to repeatedly collect the colorimetry data L*, a*, b* of pellet ore of different colors from the preliminarily classified pellet ore in S2 with the same recognition standard for multiple times, and take the average value of each to establish a corresponding data set of different pellet colors;

[0061] wherein, L* is the brightness value, a* is the red-green value, and b* is the yellow-blue value;

[0062] S4. Determination of the metallurgical properties of pellet ore: Perform multiple metallurgical property determinations on the pellet ore corresponding to the data set of different pellet colors in S3 after non-destructive testing, take the average value as the required value and record it to obtain a performance data set corresponding to the data set of different pellet colors;

[0063] S5. Establishment of linear regression model: Conduct collinearity test on the datasets of different pellet colors in S4, select appropriate colorimetric indexes as independent variables, match and correspond the selected colorimetric data with the metallurgical property results of the pellets, and establish a linear regression model;

[0064] S6. Performance test: Test the R 2 of the linear regression model in S5 and its significance level, and predict and identify the metallurgical properties of the new pellets based on this linear regression model;

[0065] S7. Transformation of the dataset and establishment of a new linear regression model: After changing the ore blending scheme and process regime, repeat the above S1 - S6 process and establish a linear regression model for the new process.

[0066] In particular, the surface pretreatment method in S1 is as follows:

[0067] (1) Wash the powder on the surface of each pellet with distilled water to remove the impurities and dust adhering to the ore surface;

[0068] (2) Put the washed pellet ore into a drying oven to dry it, thereby removing the surface moisture.

[0069] In particular, the drying in S1 is carried out in a drying oven at 105 °C for 3 h.

[0070] In particular, removing the impurities and dust adhering to the ore surface in S1 mainly refers to removing the red hematite ore powder adhering to the surface of the blue pellets.

[0071] In particular, the manual pre - classification and marking in S2 are mainly carried out by using Arabic numerals or letters or a combination of letters and numbers to distinguish and mark according to the different colors of the pellet ore surface, so as to facilitate the chromaticity - property matching and type distinction in the subsequent experimental process.

[0072] In particular, in S3, when the L* value is 0, it represents black, and when it is 100, it represents white; the a* value represents the degree of red - green. When the a* value is positive, it means the color is more red, and when the a* value is negative, it means the color is more green; the b* value represents the degree of yellow - blue. When the b* value is positive, it means the color is more yellow, and when the b* value is negative, it means the color is more blue.

[0073] In particular, the measurement methods of the colorimetric data L*, a*, b* of the pellet ore of different colors in S3 are as follows:

[0074] (1) Start the colorimeter, use the standard background tool, and calibrate the colorimeter twice under white and black backgrounds;

[0075] (2) Place the pellet ore in the colorimetric measurement area of the colorimeter. The colorimeter emits light by itself. After the light shines on the surface of the pellet ore, it is reflected, thereby obtaining the target parameters L*, a*, b*;

[0076] (3) To comprehensively consider the overall chromaticity index and performance of pellet ore and avoid contingency, each pellet ore is randomly rotated under the chromaticity meter to select five positions and measure the chromaticity twice at each position. The average of the ten obtained chromaticity values is taken as the final colorimetric index of the pellet.

[0077] (4) Due to the non-destructive nature of chromaticity meter measurement, each measured pellet ore can be individually numbered and stored classified.

[0078] Specifically, the chromaticity meter parameters measured in S3 are uniformly set to the SCI measurement mode, the D65 light source is applied, and a 10° field of view angle is selected for observation.

[0079] Specifically, the properties measured in S4 for metallurgical performance include pellet compressive strength, reducibility, and reduction swelling.

[0080] Specifically, for the establishment of the linear regression model in S5, appropriate one or more independent variables need to be selected from the previously measured colorimetric data of the chromaticity indices L*, a*, and b*, and matched with the metallurgical performance results of the pellet ore.

[0081] Specifically, there is a collinearity expansion coefficient VIF detected among L*, a*, and b* in S5, which proves that the collinearity degree of the dependent variables is serious. A linear regression equation cannot be established with these values, and other optimal multiple independent variables need to be selected.

[0082] Specifically, the performance test in S6 is as follows: Using the chromaticity index of the pellet ore as the independent variable and the metallurgical performance as the dependent variable, a linear regression equation is established, and the obtained R 2 is evaluated with the significance level. R 2 represents the goodness of fit. The significance level represents the probability of a small-probability event occurring in a single experiment. It is necessary to evaluate whether it meets the requirements according to the actual situation and production requirements, so as to predict the metallurgical performance of the pellet ore.

[0083] Example 1

[0084] A method for predicting the metallurgical performance of pellet ore based on the principle of colorimetry, the method for predicting the metallurgical performance of pellet ore based on the principle of colorimetry is as follows:

[0085] S1. Surface pretreatment of pellet ore: The selected pellet ore is subjected to surface pretreatment to obtain pellet ore with a clean surface; the method of surface pretreatment is:

[0086] (1) Wash the powder on the surface of each pellet with distilled water to remove the impurities and dust attached to the ore surface; removing the impurities and dust attached to the ore surface mainly removes the red hematite ore powder attached to the surface of the blue pellet ore.

[0087] (2) Put the cleaned pellet ore into a drying oven for drying. The drying is carried out in a drying oven at 105 °C for 3 h to remove the surface moisture;

[0088] S2. Pre-classification of pellet ore: Manually pre-classify and mark the pellet ore with clean surfaces in S1 to obtain pre-classified pellet ore; Manual pre-classification and marking are mainly carried out by distinguishing and marking with Arabic numerals according to the different surface colors of the pellet ore, so as to facilitate chromaticity-performance matching and type distinction in the subsequent experimental process; Figure 2 Pellet ore samples of different colors for chromaticity measurement;

[0089] S3. Data collection of pellet ore colorimetry: Use a colorimeter to repeatedly collect chromaticity data L*, a*, b* of pellet ore of different colors in S2 with the same recognition standard for multiple times, and take the average value of each to establish a corresponding data set of different pellet colors. The chromaticity parameters of the pellet ore are shown in Table 1 below:

[0090] Table 1 Chromaticity parameters of pellet ore in Example 1

[0091]

[0092]

[0093] Among them, L* is the brightness value, a* is the red-green value, and b* is the yellow-blue value; When the L* value is 0, it represents black, and when it is 100, it represents white; The a* value represents the degree of red and green. When the a* value is positive, it means the color is more red, and when the a* value is negative, it means the color is more green; The b* value represents the degree of yellow and blue. When the b* value is positive, it means the color is more yellow, and when the b* value is negative, it means the color is more blue;

[0094] The measurement methods of chromaticity data L*, a*, b* of pellet ore of different colors are as follows:

[0095] (1) Start the colorimeter and use the standard background tool to calibrate the colorimeter twice under white and black backgrounds;

[0096] (2) Place the pellet ore in the chromaticity measurement area of the colorimeter. The colorimeter emits light by itself, and after the light shines on the surface of the pellet ore, it is reflected, so as to obtain the target parameters L*, a*, b*;

[0097] (3) To comprehensively consider the overall chromaticity index and performance of the pellet ore and avoid contingency, rotate each pellet ore randomly under the colorimeter to select five positions and measure the chromaticity twice at each position. Take the average of the ten chromaticity values obtained as the final chromaticity index of the pellet;

[0098] (4) Due to the non-destructive nature of the colorimeter measurement, each measured pellet can be individually numbered and classified for storage;

[0099] The measured chroma meter parameters are uniformly set to the SCI measurement mode, the D65 light source is applied, and a 10° field of view angle is selected for observation;

[0100] S4. Determination of the metallurgical properties of pellet ore: The pellet ore corresponding to the data sets of different pellet colors after non-destructive testing in S3 is subjected to multiple determinations of the metallurgical properties of pellet compressive strength, and the average value is taken as the required value and recorded to obtain the performance data sets corresponding to the data sets of different pellet colors;

[0101] S5. Establishment of a linear regression model: The data sets of different pellet colors in S4 are subjected to collinearity test. It is necessary to select one or more appropriate independent variables from the previously measured colorimetric data, namely the colorimetric indexes L*, a*, b*, and match them with the results of the metallurgical properties of pellet ore to establish a linear regression model;

[0102] Before establishing a multiple linear regression equation, it is necessary to conduct a collinearity statistic test on the multiple independent variables of the established equation. The VIF value is the inflation coefficient, that is, the severity of the collinearity of the dependent variable. Generally speaking, the larger the VIF value, the more serious the collinearity. For example, if VIF > 5 or > 10, it indicates that there is a serious common relationship, and a multiple linear regression equation cannot be established with these values as independent variables.

[0103] In this study, the VIF values of L*, a*, b* and L*, a* are tested in order to select the best multiple independent variables. Table 2 shows the VIF test results of Example 1 of pellet ore;

[0104] Table 2 VIF test results of each index of pellet ore based on different independent variables in Example 1

[0105] L*, a*, b* VIF L*, a* VIF L* 4.63 L* 3.708 a* 42.05 a* 3.708 b* 50.41 / /

[0106] It can be seen from the VIF test that if L*, a*, b* are used as independent variables to establish a linear regression equation, its VIF value far exceeds the specified standard, reaching a maximum of 50.41, indicating that the linear relationship between the independent variables is very significant; while using L*, a* as independent variables to establish a linear regression equation, its VIF values do not exceed 5, indicating that there is a very strong linear relationship between a* and b*, and there is no linear relationship between L* and a*. Using L* and a* is more in line with the requirements of multiple linear regression.

[0107] According to the significant degree of the influence of the independent variable on the dependent variable, taking the pellet compressive strength F index as the dependent variable and the L* value and a* value as the independent variables, the initial parameters obtained from the previous experiments are selected to fit the regression equation F. The linear regression equation F of Example 1 is as follows:

[0108] F = 1316.71 - 11.99×L* - 289.50×a* (fitting equation of Example 1)

[0109] S6. Performance inspection: Taking the true compressive strength of the experimental results as the X-axis and the strength value predicted by linear regression as the Y-axis, a control curve is established to conduct the difference analysis of the linear regression equation, as Figure 3 .

[0110] As can be seen from the figure, the true compressive strength and the fitted compressive strength of Example 1 are basically evenly distributed on both sides of the line y = x, which indicates that the fitting result can basically well reflect the real situation. And the R 2 value is 0.794. For the P value calculated by the system, it is 2.45×10 -8 <0.001, indicating that the linear model is significant at the 0.001 level. This linear regression equation has statistical significance, and the compressive strength of the pellet can be predicted according to the chromaticity index of the pellet.

[0111] Example 2

[0112] A method for predicting the metallurgical properties of pellets based on the principle of chromaticity, the method for predicting the metallurgical properties of pellets based on the principle of chromaticity comprises the following steps:

[0113] S1. Surface pretreatment of pellets: The selected pellets are subjected to surface pretreatment to obtain pellets with a clean surface; the method of surface pretreatment is:

[0114] (1) Wash the powder on the surface of each pellet with distilled water to remove the impurities and dust attached to the surface of the ore; removing the impurities and dust attached to the surface of the ore is mainly to remove the red hematite ore powder attached to the surface of the blue pellets;

[0115] (2) Put the washed pellets into a drying oven for drying. The drying is carried out in a drying oven at 105°C for 3 hours to remove the surface moisture;

[0116] S2. Preliminary classification of pellets: The pellets with a clean surface in S1 are subjected to manual preliminary classification and marking to obtain preliminarily classified pellets; the manual preliminary classification and marking are mainly to distinguish and mark with Arabic numerals according to the different surface colors of the pellets, so as to facilitate chromaticity-performance matching and type distinction in the subsequent experimental process;

[0117] S3. Data collection of pellet chromaticity: Using a chromaticity meter, repeatedly collect the chromaticity data L*, a*, b* of pellets with different colors from the preliminarily classified pellets in S2 with the same recognition standard, and take their respective averages to establish a corresponding data set for different pellet colors. The chromaticity parameters of the pellets are shown in Table 3 below:

[0118] Table 3 Chromaticity parameters of pellets in Example 2

[0119]

[0120]

[0121] Among them, L* is the brightness value, a* is the red-green value, and b* is the yellow-blue value; when the L* value is 0, it represents black, and when it is 100, it represents white; the a* value represents the degree of red and green. When the a* value is positive, it means the color is more red, and when the a* value is negative, it means the color is more green; the b* value represents the degree of yellow and blue. When the b* value is positive, it means the color is more yellow, and when the b* value is negative, it means the color is more blue.

[0122] The measurement methods for the colorimetric data L*, a*, and b* of pellet ores of different colors are as follows:

[0123] (1) Start the colorimeter, and use the standard background tool to calibrate the colorimeter twice under the white background and the black background;

[0124] (2) Place the pellet ore in the colorimetric measurement area of the colorimeter. The colorimeter emits light by itself. After the light shines on the surface of the pellet ore, it is reflected, so as to obtain the target parameters L*, a*, and b*;

[0125] (3) To comprehensively consider the overall colorimetric index and performance of the pellet ore and avoid contingency, randomly rotate each pellet ore at five positions under the colorimeter and measure the colorimetry twice at each position. Take the average of the ten colorimetric values obtained as the final colorimetric index of the pellet ore;

[0126] (4) Due to the non-destructive nature of the colorimeter measurement, each measured pellet ore can be individually numbered and stored classified;

[0127] The parameters of the colorimeter for measurement are uniformly set to the SCI measurement mode, apply the D65 light source, and select a 10° field of view for observation;

[0128] S4. Determination of the metallurgical properties of pellet ores: Conduct multiple measurements of the metallurgical properties of the compressive strength of pellet ores corresponding to the data sets of different pellet colors after non-destructive testing in S3. Take the average value as the required value and record it to obtain the performance data sets corresponding to the data sets of different pellet colors;

[0129] S5. Establishment of a linear regression model: Conduct a collinearity test on the data sets of different pellet colors in S4. It is necessary to select one or more appropriate independent variables from the previously measured colorimetric data, namely the colorimetric indexes L*, a*, and b*, and match them with the results of the metallurgical properties of the pellet ore to establish a linear regression model;

[0130] Before establishing a multiple linear regression equation, it is necessary to conduct a collinearity statistic test on multiple independent variables of the established equation. The VIF value is the inflation coefficient, that is, the severity of the collinearity of the dependent variable. Generally speaking, the larger the VIF value, the more serious the collinearity. For example, if VIF > 5 or > 10, it indicates a serious common relationship, and a multiple linear regression equation cannot be established with these values as independent variables.

[0131] In this study, the VIF values of L*, a*, b* and L*, a* were tested to select the best multiple independent variables. Table 4 shows the VIF test results of Pellet Example 1;

[0132] Table 4 VIF Test Results of Each Index of Pellet Based on Different Independent Variables in Example 2

[0133] L*, a*, b* VIF L*, a* VIF L* 1.24 L* 1.207 a* 91.25 a* 1.207 b* 89.50 / /

[0134] It can be seen from the VIF test that if a linear regression equation is established with L*, a*, b* as independent variables, its VIF value far exceeds the specified standard, reaching a maximum of 91.25, indicating that the linear relationship between its independent variables is very significant; while using L*, a* as independent variables to establish a linear regression equation, its VIF values do not exceed 5, indicating that there is a very strong linear relationship between a* and b*, and there is no linear relationship between L* and a*. Using L* and a* is more in line with the requirements of multiple linear regression.

[0135] According to the significant degree of the influence of independent variables on the dependent variable, taking the pellet compressive strength F index as the dependent variable and the L* value and a* value as independent variables, the regression equation F was fitted with the initial parameters obtained from the previous experiments. The linear regression equation F of Example 1 is as follows:

[0136] F = 2563.87 - 12.94×L* - 193.77×a* (Fitting Equation of Example 2)

[0137] S6. Performance test: Taking the true compressive strength of the experimental results as the X-axis and the strength value predicted by linear regression as the Y-axis to establish a control curve, so as to conduct a differential analysis of the linear regression equation, such as Figure 4 .

[0138] It can be seen from the figure that the true compressive strength and the fitted compressive strength of Example 2 are basically evenly distributed on both sides of the y = x line, indicating that the fitting results can basically well reflect the real situation. And after calculation and adjustment, the R 2 value is 0.823. For the P value calculated by the system, it is 4.89×10 -9 <0.001, indicating that the linear model is significant at the 0.001 level. This linear regression equation has statistical significance, and the compressive strength of the pellet can be predicted according to the chromaticity index of the pellet.

[0139] For the above solution, the present invention provides a method for predicting the metallurgical properties of pellet ore based on the principle of colorimetry, which can solve the technical problems existing in the evaluation of the metallurgical properties of pellet ore, such as high prediction cost, low efficiency, many influencing factors in the prediction process, great operation difficulty in the prediction process, and prediction models of intelligent algorithms based on the ore phase patterns, ore phase compositions, and SEM images of pellet ore that require a large amount of measurement data.

[0140] Through surface pretreatment, the present invention can avoid the problems that different - colored pellet ores are squeezed, rubbed, and powdered against each other during transportation and cover the surfaces of other pellet ores, affecting the observation and measurement of their surface colors. Therefore, during the testing process, the intrinsic color data of pellet ore is obtained through surface pretreatment to eliminate the interference of mutual friction and powdering on color observation and measurement.

[0141] Through the pre - classification of pellet ore, the present invention can pre - distinguish and label different - colored pellet ores after preliminary treatment, which is convenient for type distinction and chromaticity - property matching in subsequent experimental processes.

[0142] Through the data acquisition of the colorimetry of pellet ore, the color of pellet ore can be quantitatively characterized by three colorimetry indexes L*, a*, and b*. Finally, the relationship between them and the metallurgical properties is established, and their correlation is tested to obtain a linear regression prediction model.

[0143] Through the determination of the metallurgical properties of pellet ore, the colorimetry indexes of different - colored pellet ores can be matched with the metallurgical properties (such as compressive strength, reducibility, reduction swelling property). According to the metallurgical property difference system under different chromaticity indexes, a performance evaluation prediction model based on colorimetry is built to solve the disadvantages of complex and time - consuming on - site detection, and promote safe production and high - efficiency intelligent production.

[0144] Through the establishment of a linear regression model, the present invention can match the colorimetry indexes of finished pellet ore with its various metallurgical properties and establish a correlation, build a metallurgical property evaluation system of pellet ore based on colorimetry, and then predict its metallurgical properties through chromaticity indexes and regression equations, helping iron and steel enterprises improve production efficiency in the performance detection of pellet ore.

[0145] Through performance inspection, the present invention can compare the true metallurgical properties of finished pellet ore with the metallurgical properties of finished pellet ore predicted based on the principle of colorimetry, test the degree of deviation, and evaluate whether it meets the requirements according to the actual situation and production requirements, so as to predict the metallurgical properties of pellet ore.

[0146] The linear regression model prediction and identification of the metallurgical properties of pellet ore designed and obtained in the present invention, which conforms to the Lab colorimetry numerical relationship, has shown significant innovation and improvement in the field of pellet ore detection and analysis, and has a positive effect on cost reduction, quality improvement and efficiency increase of iron and steel enterprises, and promoting the optimization of detection technology.

[0147] The method of the present invention not only solves the disadvantages such as danger and time-consuming in traditional metallurgical property measurement, but also is of great significance for promoting the progress of pellet ore detection and analysis means, safe production and high-efficiency intelligent production; in addition, it can provide technical support for the optimization of the process system, quality improvement in the pellet ore production process and the further development of the prospect goal of large-scale pellet ore smelting in blast furnaces.

[0148] In summary, compared with other traditional methods, the method of the present invention collects colorimetry parameters and measures the metallurgical properties of pellet ore, establishes a linear regression equation model and predicts its metallurgical properties accordingly, solves the disadvantages such as danger and time-consuming in traditional metallurgical property measurement, and is of great significance for promoting the progress of pellet ore detection and analysis means, safe production and high-efficiency intelligent production; the method is simple and easy to operate, green and environmentally friendly, with low cost, short process and high efficiency, and is conducive to large-scale industrial production and promotion.

[0149] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Among them, A and B can be singular or plural. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context before and after.

[0150] In the present invention, "at least one" means one or more, and "a plurality" means two or more. "At least one (item)" or its similar expression below refers to any combination of these items, including any combination of single item (item) or plural items (items). For example, at least one (item) of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0151] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0152] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims described above.

Claims

1. A method for predicting the metallurgical properties of pellets based on the principle of colorimetry, characterized in that: The method for predicting the metallurgical properties of pellets based on the principle of colorimetry has the following steps: S1. Surface pretreatment of pellets: Surface pretreatment of the selected pellets to obtain pellets with clean surfaces; S2. Pre-classification of pellets: artificially pre-classify and mark the pellets with cleaned surfaces in S1 to obtain pre-classified pellets; S3. Data collection of pellet colorimetry: Use a colorimeter to repeatedly collect the colorimetric data L*, a*, b* of pellets of different colors using the same identification standard for the pre-classified pellets in S2, and take their respective average values ​​to establish corresponding data sets of different pellet colors; Among them, L* is the brightness value, a* is the red-green value, and b* is the yellow-blue value; S4. Determination of metallurgical properties of pellets: perform multiple metallurgical property determinations on the pellets corresponding to the data sets of different pellet colors in S3 after nondestructive testing, take the average value as the required value and record it, and obtain the performance data sets corresponding to the data sets of different pellet colors; S5. Establishment of linear regression model: Perform collinearity test on the data sets of different pellet colors in S4, select appropriate colorimetric indicators as independent variables, match the selected colorimetric data with the metallurgical performance results of pellet ore, and establish a linear regression model; S6. Performance test: Test S5 linear regression model R 2 The linear regression model is used to predict and identify the metallurgical properties of new pellets. S7. Transformation of data set and establishment of new linear regression model: After changing the ore blending plan and process system, the above S1-S6 process should be repeated and a linear regression model of the new process should be established.

2. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: The surface pretreatment method in S1 is: (1) Wash the powder on the surface of each pellet with distilled water to remove impurities and dust attached to the surface of the ore; (2) The cleaned pellets are placed in a drying oven for drying to remove surface moisture.

3. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 2, characterized in that: The impurities and dust removed from the ore surface in S1 are mainly the red hematite powder attached to the surface of the blue pellet ore.

4. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: The manual pre-classification and labeling in S2 is mainly based on the different surface colors of the pellets, and Arabic numerals or letters or a combination of letters and numbers are used to distinguish and mark them, so as to facilitate colorimetric-performance matching and type differentiation in subsequent experiments.

5. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: In S3, an L* value of 0 represents black and a value of 100 represents white; an a* value represents the degree of redness and greenness, a positive a* value represents a reddish color, and a negative a* value represents a greenish color; a b* value represents the degree of yellowness and blueness, a positive b* value represents a yellowish color, and a negative b* value represents a bluish color.

6. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: The measurement method of the chromaticity data L*, a*, b* of pellets of different colors in S3 is: (1) Start the colorimeter and use the standard background tool to calibrate the colorimeter twice under a white background and a black background; (2) The pellets are placed in the colorimeter's colorimetric measurement area. The colorimeter emits a light source internally, and the light is reflected after hitting the pellet surface, thereby obtaining the target parameters L*, a*, and b*. (3) In order to comprehensively consider the overall color index and performance of the pellets and avoid contingency, each pellet was randomly rotated and selected at five positions under the colorimeter to measure the color twice, and the average of the ten color values ​​obtained was taken as the final color index of the pellet; (4) Due to the non-destructive nature of the colorimeter measurement, each pellet can be individually numbered and stored in a classified manner after measurement.

7. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: The colorimeter parameters measured in S3 were uniformly set to SCI measurement mode, using D65 light source and a 10° viewing angle for observation.

8. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: The properties measured in metallurgical properties of S4 include pellet compressive strength, reducibility, and reduction expansion.

9. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: In S5, there is a collinear expansion coefficient VIF between L*, a*, and b*, which proves that the dependent variables are seriously collinear. It is impossible to establish a linear regression equation with these values, and other optimal multivariate independent variables need to be selected.

10. The method for predicting metallurgical properties of pellets based on the principle of colorimetry according to claim 1, characterized in that: The performance test in S6 is as follows: taking the pellet chromaticity index as the independent variable and the metallurgical performance as the dependent variable, a linear regression equation is established and the obtained R 2 The significance level was evaluated, R 2 It represents the goodness of fit, and the significance level represents the possibility of a small probability event occurring in a test. It is necessary to evaluate whether it meets the requirements based on the actual situation and production requirements, so as to predict the metallurgical properties of pellets.

Images