JP2021165340A - Method for predicting average particle size of coke - Google Patents

Abstract

To provide a method for predicting an average particle size of coke that can estimate at least the average particle size of the coke after impact simulating transportation to a blast furnace without performing a complicated drum test or the like.SOLUTION: An area of a crack surface of the coke is calculated for each of a plurality of cokes of different types, and then a relation between an area of the crack surface of each coke and the average particle size of the coke after receiving an impact simulating transport to the blast furnace measured for each coke is obtained in advance. The area of the crack surface is calculated about the coke obtained by carbonization in a test coke oven as coke to be used in the blast furnace, and the average particle size of cokes after being impacted is predicted from the above-mentioned relation.SELECTED DRAWING: Figure 5

Description

The present invention relates to a prediction method for predicting the average particle size of coke.

The coke discharged from the coke oven receives an impact in the transport process from falling from the coke oven to reaching the blast furnace, so that the average particle size decreases. Since the average particle size of coke affects the air permeability of the blast furnace, it is an important issue in blast furnace operation to predict the average particle size of coke, particularly the average particle size after impact.

As a method of predicting the average particle size of coke after impact, an impact simulating the transport process (hereinafter, may be referred to as rotational impact) is applied by rotating the coke a specified number of times (for example, 30 rotations) in a rotating drum. After that, a drum test for sieving is known. However, the drum test has a problem that it takes a lot of labor and the test time becomes long. Prior to the drum test, a measurement test for measuring the average particle size of coke before the rotational impact is carried out. In this measurement test (hereinafter, sometimes referred to as an initial particle size measurement test), a coke cake that has been carbonized in a test coke oven is dropped from a specified height (about 2 m) to break it, and then the average particle size is determined by sieving. This is a measurement test, which is different from the drum test that simulates the impact received during the transportation process.

In Patent Document 1, each crack generated inside a coke cake obtained by drying and distilling compound coal in a test coke oven is separated into a component in the furnace width direction and a component in the furnace height direction, respectively, and the wall surface of the test coke oven. A technique for estimating the extrudability of coke cake based on the total amount (area) of crack components in the furnace height direction, which is oriented parallel to the above, is described.

In Patent Document 2, after imaging a coke cake dry-distilled in a test coke oven by X-ray CT, the cross-sectional area of the head (so-called potash flower) of a coke mass is calculated, and coke grains are calculated based on the calculated cross-sectional area. A technique for predicting the diameter is described.

Patent Document 3 describes a technique of measuring the length (total length) of cracks inside a coke breeze by X-ray CT and controlling the heating pattern of a coke oven based on the measurement result.

Patent No. 5011839 Japanese Unexamined Patent Publication No. 5-223725 Japanese Unexamined Patent Publication No. 5-230462

Patent Document 1 discloses a technique of using the area of cracks when estimating the extrudability of coke cake, and does not disclose a technique of predicting the average particle size of coke.

Further, Patent Documents 2 and 3 do not disclose a technique for predicting the average particle size of coke after impact.

An object of the present invention is to estimate at least the average particle size of coke after impact simulating transfer to a blast furnace without performing a complicated drum test or the like. Here, since it says "at least after impact", it includes not only the case of estimating only the average particle size of coke after impact but also the case of estimating the average particle size of coke before and after impact.

In order to solve the above problems, the method for predicting the average particle size of coke according to the present invention is as follows: (1) The coke in the dry coke container that has been dried in the test coke oven is imaged by X-ray CT. The types are different from the first step of acquiring information on the crack shape of coke and the second step of acquiring the crack surface and calculating the area of the crack surface by image analysis of the acquired crack shape. After calculating the area of the crack surface of each coke by executing the first step and the second step for each of the plurality of cokes, the area of the crack surface of each coke and at least the measured coke for each coke are transferred to at least the blast furnace. For the third step of determining the relationship with the average particle size of coke after receiving an impact simulating transportation, and for coke obtained by drying and distilling in a test coke furnace as coke to be used in a blast furnace. From the area of the crack surface obtained by performing the first step and the second step and the relationship obtained in the third step, at least the average particle size of coke after being impacted is predicted. It is characterized by having 4 steps.

(2) In the third step, the crack surface of each coke is divided according to the size of the crack width, and the area of the crack surface of each coke has an effect on the average particle size of the coke in the smaller crack width category. Weighting is performed so that the degree is large, and as coke to be used in the blast furnace in the fourth step, the crack surface of the coke obtained by carbonization in the test coke furnace is determined according to the above classification. The above (1) is characterized in that at least the average particle size of coke after being impacted is predicted from the area of the crack surface belonging to each category and the relationship obtained in the third step. The method for predicting the average particle size of coke described.

(3) At least the average particle size of the coke after the impact is the average particle size of the coke after the impact, or the average particle size of the coke after the impact and the coke before the impact. The method for predicting the average particle size of coke according to the above (1) or (2), which comprises both of the average particle size of the coke.

According to the present invention, it is possible to estimate at least the average particle size of coke after impact simulating transportation to a blast furnace without performing a complicated drum test or the like.

It is a transition diagram which shows the method of image processing schematically. It is a figure for demonstrating the division of a crack surface according to a crack width. This is the particle size distribution of coke breeze obtained in an actual coke oven. It is a graph which showed the correlation degree between the estimation result (MS0rev / + 25mm (estimation)) and the test result of the initial particle size measurement test (MS0rev / + 25mm) (reference). It is a graph which showed the degree of correlation between the estimation result (MS30rev / + 25mm (estimation)) and the test result of the rotational impact test (MS30rev / + 25mm) (Example 1). It is a graph which showed the degree of correlation between the estimation result (ΔMS (estimation)) and the test result (ΔMS) (Example 1). It is a graph for demonstrating the correlation between the estimation result (MS0rev / + 25mm (estimation)) and the test result of the initial particle size measurement test (MS0rev / + 25mm) (comparative example). It is a graph which showed the correlation degree between the estimation result (MS0rev / + 25mm (estimation)) and the test result of the initial particle size measurement test (MS0rev / + 25mm) (reference). It is a graph which showed the degree of correlation between the estimation result (MS30rev / + 25mm (estimation)) and the test result of the rotational impact test (MS30rev / + 25mm) (Example 2). It is a graph which showed the degree of correlation between the estimation result (ΔMS (estimation)) and the test result (ΔMS) (Example 2).

A prediction method for predicting the average particle size of coke, which is an embodiment of the present invention, comprises the following steps 1 to 4.
(Step S1)
The coke cake dried in the test coke oven is imaged by, for example, a three-dimensional medical X-ray CT before being divided into coke lumps (that is, in a dry distillation container). Information on the crack shape of the coke cake can be obtained three-dimensionally by imaging with a three-dimensional medical X-ray CT.

(Step S2)
Next, the crack surface acquired in step S1 is image-analyzed to acquire the crack surface and the area of the crack surface is calculated. A known method can be used for image analysis. For example, using the three-dimensional image analysis software Aviso, image analysis can be performed by the procedure of steps S2-1 to 2-4 below.

(About step S2-1)
In step S2-1, the captured image is subjected to image processing (for example, binarization processing based on the difference in luminance value) to separate the coke portion and the crack (space) portion. In the binarization process, for example, the coke portion can be separated into black and the crack portion can be separated into white. However, the color coding method is not limited to this, and other colors may be used. FIG. 1A is a two-dimensional schematic representation of the separated coke and cracks, but in reality, three-dimensional image data can be obtained. Note that (b) to 1 (d) of FIG. 1 cited in the following description are also two-dimensional schematic views as in FIG. 1 (a), but are actually three-dimensional image data. Is obtained.

(About step S2-2)
In step S2-2, a separation process is performed to further separate the coke portion into lumps. Here, as shown in FIG. 1B, the coke lump A looks like a single lump at first glance, but has a crack x inside and gives an impact such as a drop impact or a rotational impact. And crack x as the starting point. Therefore, even if the coke lump looks like one lump at first glance, the coke lump having the crack x is regarded as being divided when it receives an impact, and is discriminated as a different lump. In the separation process in this step, a Watershed algorithm (Non-patent document: Beucher and Meyer. The morphological approach to segmentation: the watershed transformation. Mathematical morphology in) is performed to identify each particle in the image as a separate object. Image processing; see 34, 433-81 (1993).) Can be used.

Here, when the width of the crack x is extremely small, the coke mass is not divided even if it receives an impact. Therefore, in order to consider only the cracks that affect the division at the time of impact, image processing is performed to fill the cracks below the threshold value with the width that does not contribute to the division at the time of impact as a threshold value. The width that does not contribute to the division at the time of rotational impact can be obtained by experiments or the like. Here, the threshold value is set to 0.1 mm.

In the present embodiment, as a method of simulating the impact that the coke receives during transportation to the blast furnace, a rotational impact is applied to the coke, sieving and mass measurement are performed, and the particle size distribution is calculated. For the rotational impact, for example, the drum strength tester described in JIS K2151 can be used. The number of rotations of the rotational impact can be appropriately set based on the actual results of each belt conveyor scheduled to be used, but for example, it can be typically 30 rotations. Since the impact that the coke receives during the transport process changes depending on the length of the conveyor and the like, the rotation speed of the drum is not limited to 30 times. That is, it is possible to set an appropriate rotation speed that simulates the impact received in the transport process.

(Step S2-3)
Expansion processing is evenly applied to each separated coke mass, and image processing is performed to fill the cracks. As a result, the boundary portion formed between the coke lumps separated in step S2-2 is extracted (see (c) of FIG. 1).

(Step S2-4)
The boundary part extracted in step S2-3 is compared with the crack part in step S2-1, and the surface that is the boundary part and is located at the crack part is extracted as a crack surface (FIG. 1). (D)).

(Step S2-5)
The area of the crack surface extracted in step S2-4 is standardized by the volume of the entire analysis area obtained by adding the coke portion and the crack portion, and S _all is calculated.

(Step S3)
For each of the plurality of cokes of different types, steps S1 and S2 described above are executed, and S _all is calculated for each coke. Here, the different types of coke are coke having different blending conditions (coal brand, blending ratio, bulk density, etc.) and heating conditions (heating rate, reaching temperature, heating time) of the blended coal. For each coke, the average particle size of the coke before the rotational impact (in other words, the average particle size of the coke measured by the initial particle size measurement test): MS _before (measured value) and the average of the coke after the rotational impact. Grain size (in other words, average grain size of coke measured by drum test): MS _after (measured value) and particle size difference of average grain size of coke before and after rotational impact: ΔMS (measured value) is obtained in advance. Keep it. By the above-mentioned processing, a plurality of sets of data in which the measured value regarding the average particle size of coke and S _all are set as one set of data are obtained, and by fitting these data to a linear function based on the least squares method, the following Build a linear expression.

(Step S4)
Therefore, the above formulas (1) to (3) are obtained in advance, and coke measured by newly performing steps S1 to S3 (as coke to be used in a blast furnace, it is obtained by drying in a test coke oven). _{By substituting S all of the} obtained coke into these equations, the average coke particle size before the rotational impact (in other words, the average coke particle size measured by the initial particle size measurement test): MS _before (estimated). ) And the average particle size of coke after the rotational impact (in other words, the average particle size of coke measured by the drum test): MS _after (estimated) and the difference in the average particle size of the coke before and after the rotational impact: Δ MS (estimation) can be estimated.

According to the method of the present embodiment, it is possible to estimate at least the average particle size of coke after a rotational impact without performing a drum test or the like on the coke cake carbonized in a test coke oven. That is, by acquiring the three-dimensional crack shape of the coke cake and calculating the area of the crack surface, it is possible to estimate at least the average particle size of coke after the rotational impact. In this way, by estimating the average particle size of coke after the rotational impact, the impact in the transport process until reaching the blast furnace is simulated, so that it can be used for controlling the air permeability of the blast furnace.
In addition, by estimating the average particle size of coke before and after the rotational impact, it is possible to know the susceptibility of coke to crack during the transport process until it reaches the blast furnace. It may be reflected in the operation of.

(Second Embodiment)
In the present embodiment, the width of the crack is divided according to the size, and the average particle size of the coke is estimated in consideration of the degree of influence on the average particle size of the coke of the crack belonging to each category. Here, during the drum test, the coke is divided into a plurality of coke lumps starting from a relatively small crack. That is, it is considered that the division of coke due to the rotational impact is dominated by cracks having a relatively small width (hereinafter, may be referred to as narrow crack widths). Therefore, in the present embodiment, when estimating the average particle size of coke before and after the rotational impact, weighting processing is performed according to the width of the crack in addition to the area of the crack surface.

Here, the crack width is, as shown in FIG. 2A, the distance P between the coke lumps facing each other across the crack surface shown by the dotted line. In FIG. 2B, the crack surface is divided into three categories according to the size of the crack width, the category 1 is the category with the smallest crack width, and the category 3 is the category with the largest crack width. However, the number of divisions is not limited to 3, and may be 2 or more.

A new method for the model-independent assessment of thickness in three-dimensional images. Journal of Microscopy; 185 (1), 67- The evaluation method shown in 75 (1997)) can be used. Specifically, it is possible to use a method of arranging the largest sphere in contact with the boundary between the area and the outer area in the target three-dimensional area, and the width of the object is calculated from the diameter of the placed sphere. Can be measured. In the case of two dimensions, the two-dimensional crack width can be evaluated by arranging a circle instead of the sphere in three dimensions, but in the case of two dimensions, the width is different even at the same part depending on the orientation of the cross section. Three-dimensional evaluation is desirable because it may change.

ただし、ｎは亀裂幅の区分番号である。Ｊは各区分に属する亀裂の回転衝撃前のコークスの平均粒径に与える影響度を表した係数であり、上述のデータを用いた重回帰分析により算出できる。Ｓは各区分に属する亀裂の亀裂面の面積である。ω_{ｂｅｆｏｒｅ}は、定数であり、同様に上述のデータを用いた重回帰分析によって算出できる。なお、ｎ及びＳの意味は、下記の式（５）及び式（６）においても同様である。

ただし、Ｋは、各区分に属する亀裂の回転衝撃後のコークスの平均粒径に与える影響度を表した係数であり、同様に重回帰分析によって算出できる。ω_{ａｆｔｅｒ}は、定数であり、同様に重回帰分析によって算出できる。

ただし、Ｌは各区分に属する亀裂の回転衝撃前後のコークスの平均粒径の差分に対する影響度を表した係数であり、同様に重回帰分析によって算出できる。ω_{ａｆｔｅｒ}は、定数であり、同様に重回帰分析によって算出できる。 The upper limit of the narrow crack width is preferably 2.5 mm. By weighting so that the degree of influence of the division having a smaller crack width becomes large, it is possible to improve the estimation accuracy of the average particle size of coke after the rotational impact. That is, as the process corresponding to step S2 of the first embodiment, the area of the crack surface is obtained for each category. Then, as a process corresponding to step S3 of the first embodiment, for each of the different types of coke, the area of the crack surface is obtained for each category by the above-mentioned method, and the coke before and after the rotational impact of the crack belonging to each category is obtained. Weighting processing considering the degree of influence on the average particle size is performed on the area of the crack surface of each category. At this stage, the degree of influence is unknown. Then, for each coke, by adding up the areas of the cracked surfaces after weighting in each category, an equation having the degree of influence and ω described later as a coefficient (hereinafter, may be referred to as a derivation equation) is derived. Further, for each coke, the average particle size of the coke before the rotational impact: MS _before (actual measurement value) and the average particle size of the coke after the rotational impact: MS _after (actual measurement value) by the same method as in the first embodiment. ) And the particle size difference of the average particle size of coke before and after the rotational impact: ΔMS (actual measurement value) is obtained in advance. By the above-mentioned processing, a plurality of sets of data obtained by using the actually measured values and the derivation formulas regarding the average particle size of coke as one set of data can be obtained. Multiple regression analysis of these data
The following equations (4) to (6) are calculated.

However, n is a division number of the crack width. J is a coefficient representing the degree of influence on the average particle size of coke before the rotational impact of cracks belonging to each category, and can be calculated by multiple regression analysis using the above data. S is the area of the crack surface of the crack belonging to each category. ω _before is a constant and can be similarly calculated by multiple regression analysis using the above data. The meanings of n and S are the same in the following equations (5) and (6).

However, K is a coefficient representing the degree of influence on the average particle size of coke after the rotational impact of cracks belonging to each category, and can be similarly calculated by multiple regression analysis. ω _after is a constant and can be similarly calculated by multiple regression analysis.

However, L is a coefficient representing the degree of influence on the difference in the average particle size of coke before and after the rotational impact of the cracks belonging to each category, and can be similarly calculated by multiple regression analysis. ω _after is a constant and can be similarly calculated by multiple regression analysis.

The crack surfaces of each category of newly measured coke (coke obtained by carbonization in a test coke furnace as coke scheduled to be used in a blast furnace) after obtaining the above formulas (4) to (6) in advance. By substituting the area of into these equations, the average particle size of coke before the rotational impact: MS _before (estimated), the average particle size of the coke after the rotational impact: MS _after (estimated), and before and after the rotational impact. Particle size difference of average coke particle size: ΔMS (estimated) can be estimated (corresponding to step S4 of the first embodiment).

According to the method of the present embodiment, it is possible to estimate at least the average particle size of coke after a rotational impact without performing a drum test or the like on the coke cake carbonized in a test coke oven. Further, by considering the width of the crack, at least the accuracy of estimating the average particle size of coke after the rotational impact can be improved.

Next, the present invention will be specifically described with reference to Examples. Various coke breeze powders having different particle sizes (see Table 1 below) were added to the blended coal and carbonized in a test coke oven. The particle size of the coking coal used for the blended coal was 3 mm or less and 85% by mass. The volatile content (ΣVM) of the blended coal was 29.2% by mass, and the total expansion coefficient (ΣTD) was 113.4%. The reason for adding powdered coke is to increase the particle size of coke. As the particle size of coke increases, the air permeability of the blast furnace increases. Therefore, powdered coke may be used as a part of the raw material in the coke manufacturing process. In this example, coke breeze was added in order to match the actual conditions of blast furnace operation.

The powdered coke used was collected in a coke dry fire extinguishing facility. The particle size distribution of the obtained coke breeze is shown in FIG. The particle size of coke breeze was adjusted by sorting with a sieve. As the test coke oven, a dry distillation vessel having a furnace width (W): 400 mm, a furnace length (L): 600 mm, and a furnace height (H): 420 mm was used. The bulk density of the blended coal was 850 dry.kg / m ³ . The ultimate temperature of the test coke oven was set to 1150 ° C. and carbonization was performed for 18.5 hours.

As a process corresponding to the above-mentioned step S3, a three-dimensional image was acquired using a medical X-ray CT with each coke placed in a carbonization container after carbonization. After imaging, an initial particle size measurement test and a rotational impact test were carried out to measure the average particle size of coke before and after the rotational impact. The measured values of these average particle sizes are shown in Table 1. The contents of the initial particle size measurement test and the rotary impact test are omitted because they have been described above. In this embodiment, the rotary impact test was set to 30 rotations as an impact simulating the transfer to the blast furnace.

The imaging conditions by X-ray CT were a tube voltage of 120 kV, a tube current of 350 mA, and a slice pitch of 0.5 mm. The size of the image was 512 × 512 pixel. The image resolution was 1.132 mm / pixel. For the separation process (corresponding to step S2), the segmentation module provided by Abizo was used.

That is, as the process corresponding to step S2-1, the coke portion and the crack (space) were separated by the binarization process. As a process corresponding to step S2-2, the watershed algorithm was used to further separate the coke portion into chunks. At this time, the parameter (marker extend) that determines the strength of the separation process was set to 1 for strong separation. As a process corresponding to step S2-3, an expansion process was performed on each of the separated coke lumps, and an image process was performed to fill the cracks. As a process corresponding to step S2-4, the boundary site extracted in step S2-3 is compared with the crack site in step S2-1, and the boundary site is located at the crack site. The surface was extracted as a cracked surface. As a process corresponding to step S2-5, the volume of the entire analysis area obtained by adding the area of the crack surface to the coke portion and the crack portion (however, the region of the void between the coke and the carbonization container is (Excluded from the standardized volume) was standardized to calculate _{S all.} S _all is shown in Table 3 of the second embodiment. The above-mentioned processing was carried out for each coke (corresponding to step S3).

The average particle size of coke before and after the measured rotational impact _{and S all} calculated by the above method were fitted into a linear function based on the least squares method, and the following equations (7) to (9) were calculated.
"0rev" indicates that the number of rotations of the drum is 0 (that is, no rotation), and "30rev" indicates the number of rotations of the drum that simulates the impact in the transport process. "+25 mm" means that the average particle size is for coke having a particle size of 25 mm or more.

FIG. 5 is a graph showing the degree of correlation between the estimated result (MS30rev / + 25 mm (estimated)) and the test result of the rotational impact test (MS30rev / + 25 mm). The horizontal axis is the estimated result, and the vertical axis is the rotational impact. It is a test result of the test. FIG. 6 is a graph showing the degree of correlation between the estimation result (ΔMS (estimation)) and the test result (ΔMS), the horizontal axis is the estimation result, and the vertical axis is the test result. FIG. 4 is a graph showing the degree of correlation between the estimation result (MS0rev / + 25 mm (estimation)) and the test result of the initial particle size measurement test (MS0rev / + 25 mm) for reference, and the horizontal axis is the estimation result. Yes, the vertical axis is the measurement result of the initial particle size measurement test.

As a comparative example, the average particle size of coke before receiving a rotational impact was predicted based on the following calculation formula using the ratio V of the volume of the crack to the entire analysis region. a _{4 and} b ₄ were calculated by the method of least squares. a _{4 and} b ₄ are shown in Table 4 of Example 2, and V is shown in Table 3 of Example 2.

FIG. 7 (comparative example) is a graph for explaining the correlation between the estimation result (MS0rev / + 25 mm (estimation)) and the test result of the initial particle size measurement test (MS0rev / + 25 mm). With reference to FIGS. 5 and 7, the method of the comparative example (method of estimating from the volume of the crack) has extremely low prediction accuracy, while the method of the example (method of estimating from the area of the crack) is relatively predictive. It turned out to be highly accurate. Further, with reference to FIGS. 6 and 7, the method of the comparative example (method of estimating from the volume of the crack) has extremely low prediction accuracy, while the method of the example (method of estimating from the area of the crack) is compared. It was found that the target prediction accuracy was high. As described above, FIG. 4 is shown for reference, but it was confirmed that there is a correlation.

Ｊ_１〜Ｊ_３はそれぞれ、区分１〜区分３に属する亀裂面の回転衝撃前のコークスの平均粒径に対する影響度を表した係数である。Ｓ_１〜Ｓ_３はそれぞれ、区分１〜３に属する亀裂面の面積である。なお、Ｓ_１〜Ｓ_３を合算すると実施例１のＳ_allとなる。Ｋ_１〜Ｋ_３はそれぞれ、区分１〜区分３に属する亀裂面の回転衝撃後のコークスの平均粒径に対する影響度を表した係数である。Ｌ_１〜Ｌ_３はそれぞれ、区分１〜区分３に属する亀裂面の回転衝撃前後のコークスの平均粒径の差分に対する影響度を表した係数である。ωは係数である。Ｊ、Ｋ、Ｌ及びωは重回帰分析によって算出した。Ｓ１〜Ｓ３は、実施例１と同様の方法で亀裂形状を三次元的に取得した後、画像解析することにより取得した。

The width of the crack is divided into 3 categories (Category 1: 2.5 mm or less, Category 2: Over 2.5 mm and 7.5 mm or less, Category 3: Over 7.5 mm), and the area of the crack surface in each category is weighted. The average particle size of coke was predicted from. That is, the crack surface was divided into three categories according to the size of the crack width by the method described in the embodiment, and the area of the crack surface belonging to each category was determined (corresponding to step S2). For each of the different types of coke, the area of the crack surface is determined for each category by the above method, and weighting processing is performed in consideration of the degree of influence on the average grain size of coke before and after the rotational impact of the cracks belonging to each category. It was performed on the area of the crack surface of the division. At this stage, the value of the degree of influence is unknown. Then, for each coke, the derivation formula was calculated by adding up the areas of the cracked surfaces after weighting in each category. Average particle size of coke before rotational impact: MS _before (actual measurement value), average particle size of coke after rotational impact: MS _after (actual measurement value), and particle size difference of average particle size of coke before and after rotational impact: For ΔMS (actual measurement value), the data of the first example was diverted. By the above-mentioned treatment, a plurality of sets of data were obtained in which the measured values and the derivation formulas regarding the average particle size of coke were used as one set of data. The following equations (11) to (13) were obtained by multiple regression analysis based on these data.

J _{1 to} J ₃ are coefficients representing the degree of influence on the average particle size of coke before the rotational impact of the crack surface belonging to Category 1 to Category 3, respectively. S _{1 to} S ₃ are the areas of the crack surfaces belonging to the categories 1 to 3, respectively. The sum of _{S 1 to} S _{3 gives S all in} Example 1. K _{1 to} K ₃ are coefficients representing the degree of influence on the average particle size of coke after the rotational impact of the crack surface belonging to Category 1 to Category 3, respectively. L _{1 to} L ₃ are coefficients representing the degree of influence on the difference in the average grain size of coke before and after the rotational impact of the crack surface belonging to Category 1 to Category 3, respectively. ω is a coefficient. J, K, L and ω were calculated by multiple regression analysis. S1 to S3 were obtained by three-dimensionally acquiring the crack shape by the same method as in Example 1 and then performing image analysis.

8 to 10 are graphs of the present embodiment corresponding to FIGS. 4 to 6, respectively.
By comparing and referring to FIGS. 5 and 9, it was found that the estimation accuracy of the average particle size of coke after the rotational impact is improved by considering the width of the crack. _{Here, since the coefficient K 1} representing the degree of influence of category 1 (2.5 mm or less) of MS30rev / + 25 mm (estimated) has a larger absolute value than the coefficient representing the degree of influence of other categories, it is after the rotational impact. It was confirmed that the influence on the average particle size of coke was dominated by cracks with a relatively small width. Further, by comparing and referring to FIGS. 6 and 10, it was found that the estimation accuracy of the particle size difference of the average particle size of the coke before and after the rotational impact is improved by considering the width of the crack. Note that FIG. 8 is shown for reference as in FIG. 4, but it was confirmed that there is a correlation.

In the above embodiment, both the average particle size of coke after the rotational impact and the average particle size of the coke before the rotational impact are estimated, but only the average particle size of the coke after the rotational impact is estimated. May be estimated.

Claims (3)

The first step of acquiring information on the crack shape of coke by performing imaging processing by X-ray CT on the coke in the carbonization container that has been carbonized in the test coke oven.
The second step of acquiring the crack surface and calculating the area of the crack surface by image analysis of the acquired crack shape, and
After calculating the area of the crack surface of each coke by executing the first step and the second step for each of a plurality of cokes of different types, the area of the crack surface of each coke and each coke were actually measured. At least the third step of obtaining the relationship with the average particle size of coke after receiving an impact that simulates transportation to a blast furnace, and
As coke to be used in the blast furnace, the area of the crack surface obtained by performing the first step and the second step on the coke obtained by carbonization in the test coke oven and the third coke. From the relationship obtained in the step, at least the fourth step of predicting the average particle size of coke after impact, and
A method for predicting the average particle size of coke, which comprises. In the third step, the crack surface of each coke is classified according to the size of the crack width, and the area of the crack surface of each coke has a large influence on the average particle size of the coke in the category having a smaller crack width. Weighting so that
In the fourth step, as coke to be used in the blast furnace, the crack surface is classified according to the above-mentioned classification with respect to the coke obtained by carbonization in the test coke furnace, and the crack surface belonging to each classification is classified. The method for predicting the average particle size of coke according to claim 1, wherein at least the average particle size of coke after being impacted is predicted from the area and the relationship obtained in the third step. The average particle size of the coke after at least the impact is the average particle size of the coke after the impact, or the average particle size of the coke after the impact and the average grain size of the coke before the impact. The method for predicting the average particle size of coke according to claim 1 or 2, wherein the coke has both diameters.

Images