Method for predicting granulated particles, method for producing granulated particles, device for producing granulated particles, and method for producing sintered ore

Through the particle size and composition prediction model, combined with the granulator and control device, the problem of difficulty in predicting particle size and composition concentration of granulated particles under changes in raw material matching conditions or moisture content is solved, and the accurate manufacturing of granulated particles and the finished product rate of sintered ore is achieved. promote.

CN118946672APending Publication Date: 2024-11-12JFE STEEL CORP
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202380030377.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-03-28
Filing Date
2023-03-06
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The prior art cannot predict the particle size and component concentration of granulated particles under different raw material mixing conditions or moisture content, especially when the particle size distribution of coke powder remains unchanged, and there is difficulty in inference.

Method used

The particle size prediction model and component prediction model are used to predict the particle size and specific component content of the granulated particles by inputting the mixing ratio and moisture content of the sintered raw materials, and the granulator and control device are used to realize the manufacturing of granulated particles.

Benefits of technology

Even under different raw material matching conditions or moisture content, the particle size and component concentration of granulated particles can be accurately predicted, which improves the yield rate of sintered ore.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118946672A_ABST
    Figure CN118946672A_ABST
Patent Text Reader

Abstract

Provided are a method for predicting granulated particles, a method for producing granulated particles, a device for producing granulated particles, and a method for producing sintered ore, with which it is possible to predict the particle size and component concentration of granulated particles even when the blending conditions of raw materials or the moisture content are different. This method for predicting granulated particles is a method for predicting granulated particles in which water is added to a sintering raw material containing an iron-containing raw material and coke powder and granulation is performed using a granulator, and the particle size of the granulated particles is predicted using a particle size prediction model. The particle size prediction model uses, as input, granulation conditions including the mixing proportion of the raw materials contained in the sintering raw materials and the moisture content of the sintering raw materials during granulation, and uses, as output, the particle size of the granulated particles.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a method for predicting granulated particles granulated from a sintering raw material by a granulator, a method for producing granulated particles, a granulated particle production device, and a method for producing sintered ore. Background Art

[0002] In the sintering process, the sintering raw materials including iron ore, coke powder, CaO (calcium oxide) and the like are not used directly, but are used after being granulated into granulated particles. Granulated particles are particles obtained by granulating the sintering raw materials into granules by a granulator such as a drum mixer. The particle size distribution of the granulated particles and the composition of each particle size are very important for the productivity of the sintered ore. For example, the upper layer yield of the sintered ore can be improved by segregating the upper layer of the coke powder. From the perspective of improving the yield, it is important to grasp the carbon composition of each particle size. Here, Patent Document 1 records an inference method: if there is a constant occurrence matrix, that is, the particle size distribution of the target raw material (such as coke powder) is unchanged, the composition of each particle size can be inferred.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Publication No. 2020-012185 Summary of the invention

[0006] However, it is known that even when the particle size distribution of coke powder remains unchanged, if the mixing conditions of the raw materials are changed, the composition of each particle size will change significantly. Therefore, in the technology of Patent Document 1, there is a problem that the composition of each particle size cannot be estimated when the mixing conditions of the raw materials are changed. In addition, Patent Document 1 does not consider the moisture content during granulation, and there is a problem that it cannot be estimated when the moisture content changes.

[0007] The present disclosure is completed in view of such a problem, and its purpose is to provide a method for predicting granulated particles, a method for producing granulated particles, a granulated particle production device and a method for producing sintered ore, which can predict the particle size and component concentration of the granulated particles even when the mixing conditions or moisture content of the raw materials are different.

[0008] A method for predicting granulated particles according to one embodiment of the present disclosure is a method for predicting granulated particles produced by adding water to a sintered raw material comprising an iron-containing raw material and coke powder and granulating the granulated particles using a granulator, wherein a particle size prediction model is used to predict the particle size of the granulated particles, and the particle size prediction model takes as input granulation conditions including the mixing ratio of raw materials contained in the sintered raw material and the moisture content of the sintered raw material during granulation, and takes as output the particle size of the granulated particles.

[0009] A method for producing granulated particles according to one embodiment of the present disclosure is a method for producing granulated particles by adding water to a sintering raw material including an iron-containing raw material and coke powder to perform granulation.

[0010] The method for predicting granulated particles uses the particle size prediction model to obtain setting values ​​of the granulation conditions that can achieve the target particle size of granulated particles.

[0011] The granulated particles are granulated using the set values ​​acquired by the method for predicting the granulated particles.

[0012] A granulated particle production device according to one embodiment of the present disclosure includes:

[0013] a granulator for granulating granulated particles from a sintered raw material including a plurality of raw materials; and

[0014] A control device for inferring the particle size distribution of the granulated particles granulated by the granulator;

[0015] The control device predicts the particle size of the granulated particles using a particle size prediction model, which takes as input the granulation conditions including the mixing ratio of the raw materials contained in the sintering raw materials and the moisture content of the sintering raw materials during granulation, and takes as output the particle size of the granulated particles.

[0016] A method for predicting granulated particles according to one embodiment of the present disclosure is a method for predicting granulated particles obtained by adding water to a sintering raw material including an iron-containing raw material and coke powder and granulating the granulated particles using a granulator.

[0017] Among them, a component prediction model is used to predict the content of a specific component of the above-mentioned granulated particles in the above-mentioned one particle size classification. The component prediction model takes as input the granulation conditions including the mixing ratio of the raw materials contained in the above-mentioned sintering raw materials and the moisture content of the above-mentioned sintering raw materials during granulation, and uses as output the content of a specific component of the granulated particles in one of the particle size classifications into which the particle size of the above-mentioned granulated particles is divided.

[0018] A method for producing granulated particles according to one embodiment of the present disclosure is a method for producing granulated particles by adding water to a sintering raw material including an iron-containing raw material and coke powder to perform granulation.

[0019] The method for predicting granulated particles uses the component prediction model to obtain setting values ​​of the granulation conditions that can achieve the target content of the specific component.

[0020] The granulated particles are granulated using the set values ​​acquired by the method for predicting the granulated particles.

[0021] A granulated particle production device according to one embodiment of the present disclosure includes:

[0022] a granulator for granulating granulated particles from a sintered raw material including a plurality of raw materials; and

[0023] A control device for inferring the particle size distribution of the granulated particles granulated by the granulator;

[0024] In the above control device,

[0025] The content of a specific component of the above-mentioned granulated particles in the above-mentioned one particle size classification is predicted using a component prediction model, the component prediction model taking as input the granulation conditions including the mixing ratio of the raw materials contained in the above-mentioned sintering raw materials and the moisture content of the above-mentioned sintering raw materials during granulation, and taking as output the content of a specific component of the granulated particles in one of the particle size classifications into which the particle size of the above-mentioned granulated particles is divided.

[0026] A method for producing sintered ore according to an embodiment of the present disclosure is to charge granulated particles produced by the above-mentioned method for producing granulated particles into a tray of a sintering machine to form a charge layer, and sinter the charge layer by the sintering machine to produce sintered ore.

[0027] According to the present disclosure, a method for predicting granulated particles, a method for producing granulated particles, a granulated particle production apparatus, and a method for producing sintered ore can be provided, which can predict the particle size and component concentration of the granulated particles even when the mixing conditions or moisture content of the raw materials are different. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram illustrating a granulation device for implementing a method for predicting granulated particles and a method for producing granulated particles according to an embodiment of the present disclosure.

[0029] Figure 2 It is a diagram showing a configuration example of a control device.

[0030] Figure 3 This is a graph showing the relationship between the water content (mass %) and the content (mass %) of granulated particles.

[0031] Figure 4 This is a graph showing the relationship between the water content (mass %) and the carbon concentration (mass %) of each particle size.

[0032] Figure 5 This is a diagram illustrating the correlation between the experimental value of the content (mass %) of granulated particles of "8.0-2.8" and the calculated value based on regression analysis.

[0033] Figure 6 This is a diagram illustrating the correlation between the experimental value of the carbon concentration (mass %) of each particle size of "8.0-2.8" and the calculated value based on regression analysis.

[0034] Figure 7 This is a diagram illustrating the correlation between the experimental value (mass %) of the particle size of granulated particles of "-2.8" and the calculated value (mass %) using a multiple regression model. DETAILED DESCRIPTION

[0035] Hereinafter, a method for predicting granulated particles, a method for producing granulated particles, an apparatus for producing granulated particles, and a method for producing sintered ore according to an embodiment of the present disclosure will be described with reference to the drawings.

[0036] Figure 1 It is a schematic diagram showing an example of a granulation device 1 for implementing the prediction method of granulated particles and the manufacturing method of granulated particles of the present embodiment. The granulation device 1 has a mixing tank 2 for iron-containing raw materials, a mixing tank 3 for CaO-containing raw materials, a mixing tank 4 for coke powder, a drum mixer 5 as a granulator, and a control device 10. A predetermined amount of the iron-containing raw material, the CaO-containing raw material and the coke powder are cut out from each mixing tank and mixed as a sintering raw material. Granulation water is added to the sintering raw material as moisture and granulation is performed using the drum mixer 5. Here, the drum mixer 5 is an example of a granulator, and a pelletizer can also be used instead. Here, the device including the granulator and the control device 10 in the granulation device 1 can be referred to as a granulated particle manufacturing device.

[0037] The granulated particles are transported to a sintering machine, for example, and sintered into sintered ore by the sintering machine. Figure 1 In the example, the sintering raw material is described as including the iron-containing raw material, the CaO-containing raw material and the coke powder, but the sintering raw material only needs to include at least the iron-containing raw material and the coke powder.

[0038] The control device 10 controls the cutting of the iron-containing raw material from the mixing tank 2, the cutting of the CaO-containing raw material from the mixing tank 3, the cutting of the coke powder from the mixing tank 4, the addition of water, and the operation of the drum mixer 5. In addition, the control device 10 predicts the content of the granulated particles of one of the particle size classifications into which the particle size of the granulated particles granulated by the granulation equipment 1 is divided. In addition, the control device 10 predicts the content of the specific component of the granulated particles. Here, in the present embodiment, the particle size classification is divided into three levels (greater than 8.0 mm, 2.8 mm to 8.0 mm, less than 2.8 mm). Among them, the particle size classification is not limited to three levels, and it can also be divided into more or less. In addition, the content of the granulated particles of one of the particle size classifications divided into a plurality of particle size classifications is called "particle size". In addition, in the description, tables and drawings of the present embodiment, the above three levels are sometimes expressed as "+8.0", "8.0-2.8" and "-2.8". That is, "+8.0" indicates a particle size classification larger than 8.0 mm, "8.0-2.8" indicates a particle size classification of 2.8 mm to 8.0 mm, and "-2.8" indicates a particle size classification smaller than 2.8 mm. In such a particle size classification, the particle size distribution of the granulated particles can be obtained by obtaining the particle sizes of all particle size classifications.

[0039] Figure 2 1 is a diagram showing an example of the configuration of the control device 10. The control device 10 has a processing unit 11, a storage unit 12 and an output unit 13. The control device 10 may also be a general-purpose computer such as a workstation or a personal computer. The processing unit 11 is, for example, a CPU, etc., and uses the program and data stored in the storage unit 12 to control the blending tank 2, the blending tank 3, the blending tank 4 and the drum mixer 5 to predict the content of the granulated particles and the content of the specific components of the granulated particles. The storage unit 12 is, for example, a flash memory that can update the records, an information recording medium such as a hard disk, a memory card, etc. that is built-in or connected to a data communication terminal, and a read-write device thereof. The storage unit 12 stores programs for realizing the various functions of the processing unit 11, data used in program execution, etc. The output unit 13 is, for example, an LCD or CRT display, etc.

[0040] The processing unit 11 includes a raw material information acquisition unit 111, a granulation condition acquisition unit 112, a granulation result prediction unit 113, and an instruction information acquisition unit 114. Figure 2 The arrows briefly show the flow of processing and major data.

[0041] The raw material information acquisition unit 111 acquires raw material information including the composition and particle size of the raw materials matched with the sintering raw materials, and the mixing amount of each raw material. The mixing amount of each raw material can be acquired as a mixing ratio. For example, when iron ore, limestone and coke powder are mixed as sintering raw materials, the raw material information acquisition unit 111 acquires the particle size, component composition and mixing ratio of each of the iron ore, limestone and coke powder. Here, the particle size, component composition and mixing ratio can be automatically input to the raw material information acquisition unit 111 via the network, or can be input to the raw material information acquisition unit 111 by an operator. In addition, the mixing ratio can be acquired from the setting information or actual value of the mixing tank 2, the mixing tank 3 and the mixing tank 4. The raw material information acquisition unit 111 outputs the raw material information including the composition, particle size of each raw material matched in the sintering raw material and the mixing ratio of each raw material to the granulation condition acquisition unit 112.

[0042] The granulation condition acquisition unit 112 acquires the granulation conditions from the raw material information acquisition unit 111 and the drum mixer 5. The granulation conditions include at least the mixing ratio of the raw materials contained in the sintering raw materials obtained from the raw material information acquisition unit 111 and the moisture content of the sintering raw materials during granulation. The moisture content of the sintering raw materials during granulation can be obtained based on the moisture content of each raw material and the granulation water added to the drum mixer 5. The moisture content of each raw material can be measured using an infrared moisture meter, for example. The granulation conditions may also include at least one of the particle size of the raw material, the duty cycle of the drum mixer 5, the rotation speed, and the residence time.

[0043] The granulation result prediction unit 113 predicts the particle size and the content of the specific component of the granulated particles granulated by the drum mixer 5. Here, the granulation result prediction unit 113 only needs to predict at least one of the particle size of the granulated particles and the content of the specific component. In the present embodiment, the granulation result prediction unit 113 predicts the particle size of the granulated particles and the content of the specific component, but for example, only the particle size of the granulated particles can be predicted, or only the content of the specific component of the granulated particles can be predicted.

[0044] First, the prediction of the particle size of the granulated particles is described. The granulation result prediction unit 113 predicts the particle size of the granulated particles of a particle size classification. For example, the granulation result prediction unit 113 reads a particle size prediction model that takes the particle size of the coke powder, the mixing ratio of each raw material and the moisture content of the sintered raw material during granulation as input and takes the particle size of the granulated particles as output from the storage unit 12. The granulation result prediction unit 113 can predict the particle size of the granulated particles by inputting the particle size of the coke powder, the mixing ratio of each raw material and the moisture content of the sintered raw material during granulation into the particle size prediction model.

[0045] The particle size prediction model can be a multiple regression model whose explanatory variables are the particle size of coke powder, the proportion of each raw material and the moisture content of the sintered raw material during granulation, and whose target variable is the particle size of the granulated particles. In addition, the particle size prediction model can be a trained machine learning model whose input is the particle size of coke powder, the proportion of each raw material and the moisture content of the sintered raw material during granulation, and whose output is the particle size of the granulated particles.

[0046] In the case where the particle size prediction model is a multiple regression model, a data set including the actual data of the particle size of coke powder, the mixing ratio of each raw material, the moisture content of the sintered raw material during granulation, and the particle size of the granulated particles is used to calculate the parameters of the multiple regression model in advance. The generated multiple regression model is stored in the storage unit 12. In addition, in the case where the particle size prediction model is a machine learning model, a data set including the actual data of the particle size of coke powder, the mixing ratio of each raw material, the moisture content of the sintered raw material during granulation, and the particle size of the granulated particles is used to perform machine learning in advance. The trained machine learning model is stored in the storage unit 12. The parameters of the multiple regression model and the trained machine learning model can be updated using the additional data set obtained after the generation.

[0047] An example of predicting the particle size of granulated particles is described. Using the same method, the granulation result prediction unit 113 can also predict the content of a specific component of granulated particles of a particle size classification. The granulation result prediction unit 113 uses a component prediction model. The component prediction model takes the particle size of coke powder, the mixing ratio of each raw material and the moisture content of the sintered raw material during granulation as input, and takes the content of a specific component of the granulated particles of one particle size classification among the particle size classifications into which the particle size of the granulated particles is divided as output. The component prediction model and the particle size prediction model can also be a multiple regression model or a trained machine learning model. Using a data set composed of actual data of the particle size of coke powder, the mixing ratio of each raw material, the moisture content of the sintered raw material during granulation and the content of a specific component of the granulated particles, the parameters of the multiple regression model or the trained machine learning model are calculated in advance and stored in the storage unit 12.

[0048] In the prediction method of granulated particles of the present embodiment, the granulation result prediction unit 113 performs prediction using a particle size prediction model and a component prediction model that use as input the granulation conditions including the mixing ratio of raw materials contained in the sintering raw materials and the moisture content of the sintering raw materials during granulation. Since the prediction method of granulated particles of the present embodiment uses the mixing ratio of raw materials and the moisture content of the sintering raw materials during granulation as input, it is possible to predict the particle size and the content of a specific component of granulated particles of one particle size classification even when the mixing conditions of raw materials and the moisture content during granulation are different.

[0049] In the granulation result prediction unit 113, the particle size prediction model and component prediction model for predicting granulated particles of other particle size classifications can be used in the same way to predict the particle size and the content of specific components of granulated particles of other particle size classifications. In this way, by predicting the particle size of granulated particles of all particle size classifications, the particle size distribution of granulated particles can be predicted. In addition, by predicting the content of specific components of all particle size classifications, the component concentration of specific components of granulated particles can be predicted.

[0050] The guidance information acquisition unit 114 compares the particle size and the content of the specific component of the granulated particles predicted by the granulation result prediction unit 113 with the target values ​​of the particle size and the content of the specific component of the granulated particles, and performs guidance for approaching the target values. In the guidance information acquisition unit 114, as guidance, a set value is obtained for at least one of the particle size of the coke powder, the mixing ratio of each raw material, and the moisture content of the sintering raw material during granulation so that the particle size of the granulated particles and the content of the specific component reach the predetermined target values.

[0051] Here, segregating coke powder in the upper layer of the raw material loading layer of the sintering machine is effective for improving the yield rate of sintered ore. By increasing the content of coke powder in granulated particles of a particle size that is easy to load into the upper layer, the yield rate of sintered ore can be improved. Granulated particles of a particle size that is easy to load into the upper layer are mastered by performing loading experiments. For example, based on the results of such loading experiments, the target values ​​of the particle size of the granulated particles and the content of specific components are determined.

[0052] The guidance information acquisition unit 114 can, for example, calculate the set value by back-calculating using the target value, the particle size prediction model, and the component prediction model as described above. The calculated set value is at least one of the particle size of the coke powder, the mixing ratio of each raw material, and the moisture content of the sintering raw material during granulation. For example, when the particle size of the granulated particles is obtained to be the set value of the target value, the guidance information acquisition unit 114 reads the particle size prediction model from the storage unit 12. Then, the guidance information acquisition unit 114 predicts the particle size of the granulated particles while changing the particle size of the input coke powder, the mixing ratio of each raw material, and the moisture content of the sintering raw material during granulation. The guidance information acquisition unit 114 determines whether the predicted particle size of the granulated particles reaches the target value. When the guidance information acquisition unit 114 determines that the predicted particle size of the granulated particles reaches the target value, the particle size of the input coke powder, the mixing ratio of each raw material, and the moisture content of the sintering raw material during granulation are used as the set value. The guidance information acquisition unit 114 outputs the set value thus obtained to the output unit 13.

[0053] The output unit 13 outputs the set value obtained from the guidance information acquisition unit 114 to, for example, an operator. When the output unit 13 is an LCD, the output unit 13 displays the set value. The operator can adjust the granulation conditions of the granulated particles according to the set value displayed by the output unit 13. Here, the guidance information acquisition unit 114 can automatically adjust the granulation conditions of the granulated particles using the particle size of the coke powder, the mixing ratio of each raw material, and the moisture content of the sintered raw material during granulation obtained as the set value. Using the set value obtained by the prediction method of the granulated particles of the present embodiment, the granulation equipment 1 can granulate the granulated particles. That is, the granulation equipment 1 can implement the manufacturing method of the granulated particles using the set value. Furthermore, the following manufacturing method of sintered ore can be implemented: the manufactured granulated particles are loaded into the tray of the sintering machine to form a loading layer, and the loading layer is sintered by the sintering machine to manufacture the sintered ore.

[0054] As described above, the granulated particle prediction method, granulated particle manufacturing method, granulated particle manufacturing device and sintered ore manufacturing method of this embodiment, through the above-mentioned structure, can predict the particle size of the granulated particles and the component concentration of each particle size even when the raw material mixing conditions or moisture content are different.

[0055] The embodiments of the present disclosure are described based on the accompanying drawings, but it should be noted that those skilled in the art can easily make various modifications or corrections based on the present disclosure. For example, the functions contained in each component can be reconfigured without logical contradiction, and multiple components can be combined into one or divided. The embodiments of the present disclosure can also be implemented as a program executed by a processor possessed by the device or a storage medium having the program recorded thereon. It should be understood that these are also included in the scope of the present disclosure.

[0056] For example, Figure 2 In the example of FIG. 1 , the control device 10 is described as having the guidance information acquisition unit 114, but the control device 10 may not have the guidance information acquisition unit 114. However, since the guidance information acquisition unit 114 can easily adjust the production conditions of the granulated particles for predicting the particle size and component concentration of the granulated particles, the control device 10 preferably has the guidance information acquisition unit 114.

[0057] In addition, in the above-mentioned embodiment, the particle size prediction model and the component prediction model are described by including the particle size of coke powder in the input, but the particle size of coke powder may not be included. In the particle size prediction model and the component prediction model, as input, as long as the mixing ratio of raw materials and the moisture content of sintering raw materials during granulation are included, the particle size of coke powder may not be included in the input. However, since the particle size of each raw material will affect the granulation of the sintering raw material, it is preferable to include the particle size of coke powder in the input. Here, the particle size or component concentration of other raw materials can be used in the input instead of the particle size of coke powder.

[0058] In addition, for the particle size prediction model and the component prediction model, at least one of the duty cycle of the drum mixer 5, the rotation speed of the drum mixer 5, and the residence time can be further included in the input. As the residence time, an estimated value estimated based on the processing amount of the sintering raw material of the drum mixer 5 can be used. As described in Example 2 below, the duty cycle of the drum mixer 5, the rotation speed of the drum mixer 5, and the residence time affect the particle size and component concentration of the granulated particles. Therefore, by including at least one of the duty cycle of the drum mixer 5, the rotation speed of the drum mixer 5, and the residence time in the input of the particle size prediction model and the component prediction model, the particle size of the granulated particles and the component concentration of a specific component of each particle size can be predicted with higher accuracy.

[0059] Hereinafter, the effects of the present disclosure will be specifically described based on examples, but the present disclosure is not limited to these examples.

[0060] (Example 1)

[0061] In Example 1, the test method is as follows. First, 8 kg of the sintering raw material prepared in accordance with the prescribed formula is added to a drum mixer 5 having a diameter of 0.3 m and a length of 0.4 m, and water is added to obtain a prescribed moisture content, and the sintering raw material is granulated. Then, the granulated particles are sieved using a sieve with a prescribed mesh, and the particle size distribution and the chemical composition of each particle size are measured. In this test method, the carbon concentration is measured as the chemical composition of each particle size.

[0062] Figure 3 and Figure 4 This is a graph showing the influence of the water content during granulation of T1 to T4 blended with the same sintering raw materials. Figure 3 This is a graph showing the relationship between the water content (mass %) and the content (mass %) of granulated particles. Figure 4It is a graph showing the relationship between moisture content (mass %) and carbon concentration (mass %) of each particle size. The graph and curve for moisture content of 6.8 mass % are the measured data (experimental value) of T1. The graph and curve for moisture content of 7.8 mass % are the measured data of T2. The graph and curve for moisture content of 8.8 mass % are the measured data of T3. The graph and curve for moisture content of 9.8 mass % are the measured data of T4. Figure 3 and Figure 4 It can be seen that the particle size distribution and the carbon concentration of each particle size change as the moisture content during granulation changes.

[0063] In addition, Table 1 shows the measured data on sintering raw materials of various formulations including T1 to T4. In Table 1, the proportion of each raw material when the formulation is set to 100 as a whole is expressed by a numerical value. "Ore A", "Ore B" and "Ore C" are iron ores of different grades or origins, respectively. For the particle size of coke powder, the proportion of particles less than 2 mm is expressed by "-2 mm", and the proportion less than 1 mm is expressed by "-1 mm". For the particle size distribution of granulated particles and the carbon concentration of each particle size, the content and concentration of each particle size classification mentioned above are expressed by mass %. It can be seen from Table 1 that as the formulation and coke powder particle size change, the particle size distribution and the chemical composition of each particle size change.

[0064]

[0065] Using the actual data shown in Table 1, the parameters of the multiple regression model as a particle size prediction model and the parameters of the multiple regression model as a component prediction model were obtained, and the particle size of the granulated particles and the chemical composition of each particle size were estimated using these particle size prediction model and component prediction model.

[0066] Figure 5 and Figure 6 This is a diagram illustrating the correlation between experimental values, the particle size of granulated particles estimated by regression analysis, and the chemical components of each particle size. Figure 5 This is a diagram illustrating the correlation between the experimental value of the content (mass %) of granulated particles of "8.0-2.8" and the calculated value based on regression analysis. Figure 6 This is a graph showing the correlation between the experimental value of the carbon concentration (mass %) of each particle size of "8.0-2.8" and the calculated value based on regression analysis. Figure 5 and Figure 6 As shown, there is a strong correlation, and the particle size of the granulated particles and the chemical composition of each particle size can be inferred based on the various formulations and the particle size of the coke powder.

[0067] In the above example, the particle size prediction model and the component prediction model are multiple regression models (inference formula based on multiple regression). Here, the particle size prediction model and the component prediction model can be trained machine learning models. As shown in Table 2, it can be seen that the results of using trained machine learning to obtain each target variable have a larger correlation coefficient than multiple regression, which becomes an inference method that can infer the particle size and chemical composition of granulated particles with higher accuracy than multiple regression. Here, "C&RTree" is Classification and Regression Tree, which is an analysis method of machine learning.

[0068] [Table 2]

[0069] (Table 2)

[0070]

[0071] (Example 2)

[0072] Example 2 was implemented according to the following test method. First, 8 kg of sintering raw materials formulated according to the regulations were added to a drum mixer 5 having a diameter of 0.3 m and a length of 0.4 m, and water was added to reach the specified moisture content, and the sintering raw materials were granulated. The duty cycle, drum mixer speed, and granulation time as the operating conditions of the drum mixer 5 were changed, and the particle size of the granulated particles of "-2.8" was evaluated. Here, as the granulation time, the residence time of the sintering raw materials in the drum mixer 5 is used. In the following, the term of residence time is used as the term corresponding to the granulation time. Table 3 shows the measurement results of the particle size of the granulated particles of "-2.8" under various operating conditions.

[0073] [Table 3]

[0074] (Table 3)

[0075]

[0076] Using the actual data shown in Table 3, the parameters of the multivariate regression model of the particle size prediction model with the duty factor, drum mixer speed, and residence time as input and the particle size of -2.8 mm as output were obtained, and the particle size of -2.8 mm of the granulated particles was estimated using this particle size prediction model.

[0077] Figure 7 This is a diagram illustrating the correlation between the experimental value (mass %) of the particle size of granulated particles of "-2.8" and the calculated value (mass %) using a multiple regression model.

[0078] The results confirmed that the duty cycle, drum mixer speed, and residence time have an impact on the particle size of the granulated particles. Therefore, it can be seen that by further adding at least one of the duty cycle, drum mixer speed, and residence time to the input of the particle size prediction model, the prediction accuracy of the particle size of the granulated particles can be improved.

[0079] Explanation of symbols

[0080] 1 Granulation equipment

[0081] 2 matching grooves

[0082] 3 matching slots

[0083] 4 matching slots

[0084] 5. Drum mixer

[0085] 10 Control device

[0086] 11 Processing Department

[0087] 12 Storage unit

[0088] 13 Output section

[0089] 111 Raw Materials Information Acquisition Department

[0090] 112 Granulation Condition Acquisition Department

[0091] 113 Granulation Result Prediction Department

[0092] 114 Guidance Information Acquisition Department

Claims

1. A method for predicting granulated particles, which is a method for predicting granulated particles obtained by adding water to a sintering raw material including an iron-containing raw material and coke powder and granulating the granulated particles using a granulator, The particle size of the granulated particles is predicted using a particle size prediction model which takes as input granulation conditions including a mixing ratio of raw materials contained in the sintering raw material and a moisture content of the sintering raw material during granulation and outputs the particle size of the granulated particles.

2. The method for predicting granulated particles according to claim 1, in, The particle size distribution of the granulated particles is predicted by predicting the particle sizes of the granulated particles of all particle size classifications.

3. The method for predicting granulated particles according to claim 1 or 2, in, The granulation conditions further include at least one of a duty cycle of the granulator, a rotation speed of the granulator, and a residence time.

4. The method for predicting granulated particles according to any one of claims 1 to 3, in, Using the particle size prediction model, setting values ​​of the granulation conditions that can achieve the target particle size of granulated particles are acquired.

5. A method for producing granulated particles, comprising adding water to a sintering raw material comprising an iron-containing raw material and coke powder and granulating the granulated particles. The granulated particles are granulated using the set values ​​acquired by the method for predicting granulated particles according to claim 4.

6. A granulated particle manufacturing device, comprising: a granulator for granulating granulated particles from a sintered raw material including a plurality of raw materials; and a control device for inferring the particle size distribution of the granulated particles granulated by the granulator; In the control device, The particle size of the granulated particles is predicted using a particle size prediction model, which takes granulation conditions including a mixing ratio of raw materials contained in the sintering raw material and a moisture content of the sintering raw material during granulation as input and takes the particle size of the granulated particles as output.

7. A method for predicting granulated particles, comprising adding water to a sintering raw material comprising an iron-containing raw material and coke powder and granulating the granulated particles using a granulator, A component prediction model is used to predict the content of a specific component of the granulated particles in the one particle size classification. The component prediction model takes as input the granulation conditions including the mixing ratio of the raw materials contained in the sintering raw material and the moisture content of the sintering raw material during granulation, and uses as output the content of a specific component of the granulated particles in one particle size classification into which the particle size of the granulated particles is divided into a plurality of particle size classifications.

8. The method for predicting granulated particles according to claim 7, in, The content of the specific component in the granulated particles is estimated by estimating the content of the specific component in the granulated particles in all particle size categories.

9. The method for predicting granulated particles according to claim 7 or 8, in, The granulation conditions further include at least one of a duty cycle of the granulator, a rotation speed of the granulator, and a residence time.

10. The method for predicting granulated particles according to any one of claims 7 to 9, in, Using the component prediction model, setting values ​​of the granulation conditions that can achieve the target content of the specific component are acquired.

11. A method for producing granulated particles, comprising adding water to a sintering raw material comprising an iron-containing raw material and coke powder to perform granulation, The granulated particles are granulated using the set value acquired by the method for predicting granulated particles according to claim 10.

12. A granulated particle manufacturing device comprising: a granulator for granulating granulated particles from a sintered raw material including a plurality of raw materials; and a control device for inferring the particle size distribution of the granulated particles granulated by the granulator; In the control device, A component prediction model is used to predict the content of a specific component of the granulated particles in the one particle size classification. The component prediction model takes as input the granulation conditions including the mixing ratio of the raw materials contained in the sintering raw material and the moisture content of the sintering raw material during granulation, and uses as output the content of a specific component of the granulated particles in one particle size classification into which the particle size of the granulated particles is divided into a plurality of particle size classifications.

13. A method for producing a sintered ore, comprising: charging granulated particles produced by the method for producing granulated particles according to claim 5 or 11 into a tray of a sintering machine to form a charging layer, and sintering the charging layer using the sintering machine to produce a sintered ore.

Citation Information

Patent Citations

  • Height direction raw material distribution estimation device, height direction raw material distribution estimation program, and method therefor

    JP2020012185A