Pig iron production method

Abstract

To provide a pig iron production method capable of further reducing the used amount of coke while securing the flow of hot wind in the central part of a blast furnace.SOLUTION: A pig iron production method according to an embodiment has a process in which pig iron is produced using a blast furnace having a tuyere and comprises: a step in which a first layer comprising ore raw material and a second layer comprising coke are alternately laminated inside the blast furnace; a step in which coke is charged to the central part of the blast furnace; and a step in which the ore raw material of the laminated first layer is reduced and melted while blowing auxiliary fuel into the blast furnace by hot wind from the tuyere. In the lamination step, during one charge in which the lamination unit consisting of the combination of one first layer and one second layer is laminated, the one or more charge steps are performed, and the ratio R of the mass (ton / ch) of the coke deposited on the central part with respect to the mass (ton / ch) of the ore raw material to be charged in one charge is controlled to a prescribed value α or more.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a method for producing pig iron. [Background technology]

[0002] A method for producing pig iron is known in which a first layer containing raw ore and a second layer containing coke are alternately stacked in a blast furnace, and the raw ore is reduced and melted while auxiliary fuel is blown into the blast furnace by hot air blown from a tuyere. In this process, the coke serves as a heat source for melting the raw ore, a reducing agent for the raw ore, a recarburizer for carburizing the molten iron to lower its melting point, and a spacer for ensuring gas permeability in the blast furnace. By maintaining gas permeability with the coke, the lowering of the charges charged as the first and second layers is stabilized, and the blast furnace is operated stably.

[0003] Hot air is generally blown into a blast furnace from the periphery. When the hot air reaches the center of the blast furnace, it rises through the center. If the flow of hot air rising through the center is blocked, the air permeability is likely to deteriorate. As a method for ensuring air permeability in the center, a method of charging coke intensively in the center of the blast furnace has been proposed (for example, JP 60-56003 A). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 60-56003 Summary of the Invention [Problem to be solved by the invention]

[0005] Recent CO 2Due to the demand for reducing emissions, further reduction in the amount of coke used in blast furnace operation is required. Although the above-mentioned method of concentrating coke charging in the center of the blast furnace improves permeability, a method of maintaining permeability with even less coke is desired.

[0006] The present invention has been made based on the above-mentioned circumstances, and aims to provide a method for producing pig iron that can further reduce the amount of coke used while ensuring the flow of hot air in the center of the blast furnace. [Means for solving the problem]

[0007] A method for producing pig iron according to one embodiment of the present invention is a method for producing pig iron using a blast furnace having a tuyere, and includes the steps of alternately stacking a first layer containing an ore raw material and a second layer containing coke in the blast furnace, charging coke into the center of the blast furnace, and reducing and melting the stacked ore raw material of the first layer while blowing auxiliary fuel into the blast furnace with hot air blown from the tuyere. In the stacking step, the charging step is performed one or more times during one charge in which a stacking unit consisting of one first layer and one second layer is stacked, and in the one charge, the ratio R of the mass (ton / ch) of coke deposited in the center to the mass (ton / ch) of the ore raw material charged is set to a predetermined value α or more.

[0008] In the pig iron manufacturing method, coke is charged into the center of the blast furnace for each charge of stacking the stacking unit, which is the combination of the first layer and the second layer in the stacking step, to easily ensure the flow of hot air in the center of the blast furnace. In addition, the ratio R of the mass (ton / ch) of coke accumulated in the center to the mass (ton / ch) of the ore raw material charged in each charge is set to a predetermined value α or more, thereby improving the permeability of the hot air. Therefore, even if the amount of coke used is reduced, the required permeability can be ensured, and the amount of coke can be further reduced.

[0009] The predetermined value α is preferably 0.017. By setting the predetermined value α to the above value, breathability can be easily ensured.

[0010] When the raw ore of the first layer contains iron ore pellets and a ratio of the iron ore pellets in the raw ore of the first layer is P (mass%), the predetermined value α may be calculated by the following formula 1. The required amount of coke to be charged in the center may also vary depending on the ore pile inclination angle of the first layer. In particular, since there is a certain correlation between the ratio P of the iron ore pellets in the first layer and the ore pile inclination angle of the first layer, by determining the predetermined value α in consideration of the ratio P of the iron ore pellets in the first layer, it is possible to obtain an improvement effect on permeability with high accuracy. α=0.017×(0.001×P+0.97) ···1

[0011] The strength of the coke deposited in the center should be equal to or greater than the strength of the coke contained in the second layer. From the viewpoint of permeability, it is preferable that the coke has a high strength, but on the other hand, high-strength coke is generally expensive, which leads to an increase in production costs. Therefore, by using high-strength coke only for central charging, it is possible to improve permeability while suppressing an increase in production costs.

[0012] The average particle size of the coke deposited in the center is preferably equal to or larger than the average particle size of the coke contained in the second layer. From the viewpoint of air permeability, a larger average particle size of the coke is preferable, but on the other hand, coke with a larger average particle size is generally expensive, which leads to an increase in production costs. Therefore, by using coke with a larger average particle size only for central charging, it is possible to improve air permeability while suppressing an increase in production costs.

[0013] The method includes the steps of: inputting into an artificial intelligence model, as learning data, actual values ​​of an input data group including at least the hot air temperature and blowing volume, solution loss reaction volume, furnace wall heat removal volume, residual iron volume, molten iron temperature, and the ratio R for a predetermined period from a time prior to a reference time to the reference time, and an output data group including molten iron temperature data obtained in the reducing and melting processes in the future from the reference time, and training the artificial intelligence model to predict the molten iron temperature data in the future from the reference time from the input data group; acquiring the input data group using the current time as the reference time; inputting the input data group acquired in the acquiring step into the trained artificial intelligence model using the current time as the reference time; and having the trained artificial intelligence model estimate the future temperature of the molten iron. It is preferable that the input data group acquired in the acquiring step and actual values ​​of the output data group corresponding to this input data group are used as input for the training step. In this way, the trained artificial intelligence model is used to estimate the temperature of the molten iron, and additional learning is performed using the input data group acquired in the acquisition process and the actual values ​​of the output data group corresponding to this input data group, making it possible to control the temperature of the molten iron with high accuracy based on the above ratio R.

[0014] Here, the "center" of the blast furnace refers to the region that is 0.2Z or less away from the central axis of the blast furnace, where Z is the radius of the throat. The "strength" of the coke refers to the drum strength defined in JIS-K-2151:2004. In addition, the "average particle size" refers to the arithmetic mean diameter. Effect of the Invention

[0015] As described above, by using the pig iron manufacturing method of the present invention, it is possible to further reduce the amount of coke used while ensuring the flow of hot air in the center of the blast furnace. [Brief description of the drawings]

[0016] [Figure 1] FIG. 1 is a flow diagram showing a method for producing pig iron according to one embodiment of the present invention. [Diagram 2]FIG. 2 is a schematic diagram showing the inside of a blast furnace used in the pig iron manufacturing method of FIG. [Diagram 3] FIG. 3 is a schematic enlarged partial view of the vicinity of the cohesive zone and the dripping zone in FIG. [Figure 4] FIG. 4 is a graph showing the relationship between the ratio R and the corrected K value in the embodiment. [Diagram 5] FIG. 5 is a graph showing the relationship between the coke rate and the K value in the examples. [Figure 6] FIG. 6 is a schematic diagram showing the configuration of a blast furnace charge distribution experimental device used in the examples. [Figure 7] FIG. 7 is a graph showing the relationship between the iron ore pellet ratio P (alumina ball ratio) and the ore pile inclination angle θ in the examples. [Figure 8] FIG. 8 is a graph showing the relationship between the coke ratio and the K value classified according to the proportion P of iron ore pellets in the examples. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] Hereinafter, a method for producing pig iron according to each embodiment of the present invention will be described.

[0018] [First embodiment] The pig iron manufacturing method shown in FIG. 1 is a pig iron manufacturing method for manufacturing pig iron using a blast furnace 1 shown in FIG. 2, and includes a layering step S1, a center charging step S2, and a reducing and melting step S3.

[0019] <Blast furnace> As shown in FIG. 2, the blast furnace 1 has a tuyere 1a and a tap hole 1b provided at the bottom of the furnace. Usually, a plurality of tuyere 1a are provided. The blast furnace 1 is a solid-gas counterflow type shaft furnace, in which hot air, which is high-temperature air and, as necessary, high-temperature or room-temperature oxygen, is blown into the furnace from the tuyere 1a to carry out a series of reactions such as reduction and melting of the ore raw material 11 described later, and pig iron can be taken out from the tap hole 1b. The blast furnace 1 is also equipped with a Bell-Armor type raw material charging device 2. This raw material charging device 2 will be described later.

[0020] <Lamination process> In the stacking step S1, as shown in Fig. 2, the first layers 10 and the second layers 20 are alternately stacked in the blast furnace 1. That is, the number of layers of each of the first layers 10 and the second layers 20 is two or more.

[0021] (1st layer) The first layer 10 includes a raw ore 11. The raw ore 11 is heated and reduced to molten iron F by hot air blown in from the tuyere 1a in the reduction and melting step S3.

[0022] The raw ore 11 refers to ores that are iron raw materials, and mainly contains iron ore. Examples of the raw ore 11 include fired ore (iron ore pellets, sintered ore), lump ore, carbonaceous composite agglomerated ore, metal, etc. The raw ore 11 may also contain aggregate 11a. Below, a case where aggregate 11a is mixed into the raw ore 11 will be described, but aggregate 11a is not an essential component, and the raw ore 11 does not necessarily need to contain aggregate 11a.

[0023] The aggregate 11a is intended to improve the air permeability of a cohesive zone D described later and to allow the hot air to pass to a center M of the blast furnace 1. The aggregate 11a preferably contains reduced iron molded products (HBI, hot briquette iron) obtained by compression molding reduced iron.

[0024] HBI is made by forming direct reduced iron (DRI) in a hot state. DRI has a high porosity and is disadvantageous in that it oxidizes and generates heat during marine transportation or outdoor storage, whereas HBI has a low porosity and is difficult to reoxidize. After ensuring the breathability of the first layer 10, the aggregate 11a functions as a metal and becomes molten iron. The aggregate 11a has a high metallization rate and does not require reduction, so it does not require much reducing agent to become this molten iron. Therefore, CO 2 This reduces emissions. Note that the "metallization rate" refers to the ratio of metallic iron to the total iron content [mass %].

[0025] The lower limit of the charging amount of the reduced iron casts is 100 kg, and more preferably 150 kg, per ton of pig iron. If the charging amount of the reduced iron casts is less than the lower limit, the aggregate 11a in the cohesive zone D may not be able to sufficiently ensure gas permeability in the reduction and melting step S3. On the other hand, the upper limit of the charging amount of the reduced iron casts is appropriately determined within a range in which the aggregate effect is not reduced due to an excess of aggregate, and the upper limit of the charging amount of the reduced iron casts is set to, for example, 700 kg per ton of pig iron.

[0026] The lower limit of the ratio of the average particle size of the reduced iron cast to the average particle size of the raw ore 11b excluding the aggregate 11a is preferably 1.3, more preferably 1.4. As shown in FIG. 3, even when a part of the raw ore 11b excluding the aggregate 11a in the first layer 10 melts and moves downward in the blast furnace 1 as dripping slag 12, and the raw ore 11b excluding the aggregate 11a softens and shrinks, the reduced iron cast having a high melting point does not soften. If the reduced iron cast that is larger than the raw ore 11b excluding the aggregate 11a is mixed as the aggregate 11a, the aggregate effect of the reduced iron cast is easily exerted, and the first layer 10 as a whole can be prevented from shrinking. Therefore, by setting the ratio of the average particle size to be equal to or more than the lower limit, a flow path of hot air as shown by the arrow in FIG. 3 can be secured, and therefore the air permeability in the reduction and melting process S3 can be improved. On the other hand, the upper limit of the ratio of the average particle size is preferably 10, more preferably 5. If the ratio of the average particle diameters exceeds the upper limit, it becomes difficult to uniformly mix the reduced iron casts in the first layer 10, and segregation may increase.

[0027] In addition, when the reduced iron cast contains aluminum oxide, the upper limit of the content of the aluminum oxide in the reduced iron cast is preferably 1.5% by mass, more preferably 1.3% by mass. If the content of the aluminum oxide exceeds the upper limit, the melting point of the slag may be increased and the viscosity may be increased, which may make it difficult to ensure gas permeability in the lower part of the furnace. For this reason, by setting the content of the aluminum oxide in the reduced iron cast to the upper limit or less, it is possible to prevent an increase in the amount of coke 21 used. The content of the aluminum oxide may be 0% by mass, i.e., the reduced iron cast may not contain aluminum oxide, but the lower limit of the content of the aluminum oxide is preferably 0.5% by mass. If the content of the aluminum oxide is less than the lower limit, the reduced iron cast may be expensive, which may increase the production cost of pig iron.

[0028] In addition to the raw ore 11, auxiliary materials such as limestone, dolomite, and silica stone may be charged together in the first layer 10. In addition to the raw ore 11, the first layer 10 generally uses a mixture of small coke particles obtained by sieving coke.

[0029] (2nd layer) The second layer 20 includes coke 21 .

[0030] The coke 21 serves as a heat source for melting the raw ore 11, as a generator of CO gas which is a reducing agent necessary for reducing the raw ore 11, as a recarburizer for carburizing the molten iron to lower its melting point, and as a spacer for ensuring ventilation within the blast furnace 1.

[0031] (Layering method) Various methods can be used to alternately stack the first layers 10 and the second layers 20. Here, the method will be described using as an example a blast furnace 1 equipped with a Bell-Armor type raw material charging device 2 (hereinafter, also simply referred to as "raw material charging device 2") as shown in Figure 2.

[0032] The raw material charging device 2 is provided at the top of the furnace. That is, the first layer 10 and the second layer 20 are charged from the top of the furnace. As shown in FIG. 2, the raw material charging device 2 has a bell cup 2a, a lower bell 2b, and an armor 2c.

[0033] Bell cup 2a is filled with the raw materials to be charged. When charging first layer 10, the raw materials constituting first layer 10 are charged into bell cup 2a, and when charging second layer 20, the raw materials constituting second layer 20 are charged into bell cup 2a.

[0034] Lower bell 2b is cone-shaped and spreads downward, and is disposed within bell cup 2a. Lower bell 2b can move up and down (in Fig. 2, the upward movement is shown by a solid line, and the downward movement is shown by a dashed line). When lower bell 2b moves upward, it seals the bottom of bell cup 2a, and when moved downward, a gap is formed in the extension of the side wall of bell cup 2a.

[0035] The armor 2c is provided below the lower bell 2b on the furnace wall of the blast furnace 1. When the lower bell 2b is moved downward, raw materials fall from the gap, and the armor 2c is a repulsion plate for repelling the falling raw materials. The armor 2c is configured to be able to move in and out of the blast furnace 1.

[0036] Using this raw material charging device 2, the first layer 10 can be laminated as follows. The same applies to the second layer 20. The first layer 10 and the second layer 20 are laminated alternately.

[0037] First, the lower bell 2b is positioned upward, and the raw material for the first layer 10 is charged into the bell cup 2a. When the lower bell 2b is positioned upward, the lower part of the bell cup 2a is sealed, so the raw material is filled into the bell cup 2a. The amount of material filled is the amount of each layer stacked. If the capacity of the bell cup 2a is insufficient for the amount of each layer stacked, the first layer 10 may be stacked in multiple batches. This stacking in one filling is also called "one batch."

[0038] Next, the lower bell 2b is moved downward. Then, a gap is created between the bell cup 2a and the raw materials, which fall from this gap toward the furnace wall and collide with the armor 2c. The raw materials that collide with the armor 2c and are repelled are charged into the furnace. The raw materials fall while moving toward the furnace interior due to the repulsion from the armor 2c, so they flow from the position where they fell toward the center of the furnace and accumulate. The armor 2c is configured to be able to move in and out toward the inside of the blast furnace 1, so the falling position of the raw materials can be adjusted by moving the armor 2c in and out. This adjustment allows the first layer 10 to be accumulated in a desired shape.

[0039] <Central charging process> In the central charging step S2, coke 31 is charged into the central portion M of the blast furnace 1. By charging the coke 31, a central layer 30 is formed as shown in Fig. 2. Note that the central portion M may be charged with not only the coke 31 but also a small amount of raw ore or the like.

[0040] The central charging step S2 (a step of charging) is carried out once or multiple times during one charge in which a stacking unit, which is a combination of one first layer 10 and one second layer 20, is stacked in the stacking step S1 (a step of stacking). Here, one charge means one cycle in which one first layer 10 and one second layer 20 are stacked, and for example, when the first layer 10 and the second layer 20 are processed in two batches, four processes, the first batch of the first layer 10, the second batch of the first layer 10, the first batch of the second layer 20, and the first batch of the second layer 20, are combined to form one charge.

[0041] The order of performing the central charging step S2 within one charge can be appropriately determined depending on various conditions. For example, the central charging step S2 may be performed as the first step of one charge, or may be performed in two separate steps immediately before laminating the first layer 10 and the second layer 20. In addition, when the first layer 10 and the second layer 20 are processed in two batches, the central charging step S2 may be performed in two separate steps, between the first batch of the first layer 10 and the second batch of the first layer 10, and between the second batch of the second layer 20 and the first batch of the first layer 10.

[0042] (Coke) The coke 31 may have the same properties as the coke 21 in the second layer 20, or may have different properties. When using coke with different properties, it is preferable that the strength of the coke 31 deposited in the center M is equal to or greater than the strength of the coke 21 contained in the second layer 20. From the viewpoint of air permeability, it is preferable that the coke has a high strength, but on the other hand, high-strength coke is generally expensive, which leads to an increase in production costs. For this reason, by mainly using high-strength coke for central charging, it is possible to improve air permeability while suppressing an increase in production costs. Note that the "coke deposited in the center" is mainly coke charged in the central charging process, but if coke is deposited in the center due to rolling or the like when the second layer is stacked, for example, this coke is included in the coke deposited in the center. In other words, the "coke deposited in the center" is coke deposited in the center after one charge is charged, regardless of its origin. Here, for the strength of the coke 31 deposited in the center M to be equal to or greater than the strength of the coke 21 contained in the second layer 20, this is equivalent to the coke 31 charged in the center charging process having a greater strength than the coke 21 contained in the second layer 20. When the coke 21 in the second layer 20 is deposited in the center M by rolling or the like, this coke can be considered to have the same strength as the coke 21 in the second layer 20.

[0043] In addition, it is preferable that the average particle size of the coke 31 deposited in the center portion M is equal to or larger than the average particle size of the coke 21 contained in the second layer 20. From the viewpoint of air permeability, it is preferable that the average particle size of the coke is larger, but on the other hand, coke with a large average particle size is generally expensive, which leads to an increase in production costs. For this reason, by mainly using coke with a large average particle size for center charging, it is possible to improve air permeability while suppressing an increase in production costs.

[0044] It is particularly preferable that the strength of the coke 31 deposited in the center portion M is greater than or equal to the strength of the coke 21 contained in the second layer 20, and that the average particle size of the coke 31 deposited in the center portion M is greater than or equal to the average particle size of the coke 21 contained in the second layer 20.

[0045] (Charging process) The central layer 30 can be layered by various methods, and is not particularly limited as long as the ratio R described later can be equal to or greater than a predetermined value α. For example, the central layer 30 can be layered by using a Bell-Armor type raw material charging device 2, similar to the first layer 10 and the second layer 20. Specifically, it is preferable to layer a part of the central layer 30 (a thickness equivalent to the thickness of the first layer 10 or the second layer 20 to be layered immediately after) at the center M of the blast furnace 1 using the raw material charging device 2.

[0046] In the pig iron manufacturing method, the ratio R of the mass (ton / ch) of the coke 31 deposited in the center M to the mass (ton / ch) of the raw ore 11 charged in one charge is set to a predetermined value α or more. For example, when the central charging step S2 or the like is performed multiple times in one charge, the ratio R refers to the ratio of the total amount of the coke 31 to the total amount of the raw ore 11 charged in one charge. In addition, the "mass of the raw ore to be charged" is mainly the raw ore 11 in the first layer 10 charged in the layering step S1, but also includes the raw ore in other layers if they contain raw ore. In other words, the "mass of the raw ore to be charged" is the total mass of the raw ore to be charged in one charge, regardless of its origin.

[0047] The predetermined value α can be set to 0.017. By setting the predetermined value α to the above value, the breathability can be easily ensured.

[0048] When the raw ore 11 in the first layer 10 contains iron ore pellets, the predetermined value α can be set to 0.017. However, when the proportion of the iron ore pellets in the raw ore 11 in the first layer 10 is P (mass%), the predetermined value α is preferably calculated by the following formula 1. Here, "iron ore pellets" are made from fine iron ore powder of several tens of μm, and are produced by improving the quality to have properties (e.g. size, strength, reducibility, etc.) suitable for blast furnaces. α=0.017×(0.001×P+0.97) ···1

[0049] In order to improve the gas permeability, it is desirable to form a central column of coke 31 in the central portion M of the blast furnace 1. In other words, it is preferable to charge the coke 31 in the central charging step S2 so that the central column is higher than the thickness of the first layer 10 near the center, which is determined by the charging mass of the raw ore 11 charged in one charge and the ore pile inclination angle, that is, so that the coke 31 in the central portion M protrudes from the first layer 10. The "ore pile inclination angle" refers to the angle of the inclined surface of the ore pile layer (first layer 10, etc.) from the horizontal.

[0050] The first layer 10 is generally stacked at an inclination so that the center M is lower. Here, sintered ore and lump ore are irregular and have a relatively wide particle size distribution, whereas iron ore pellets are spherical and have a relatively uniform particle size. For this reason, iron ore pellets tend to roll to the center M compared to sintered ore and lump ore. When the proportion of iron ore pellets is increased, the ore pile inclination angle tends to become smaller. At this time, the first layer 10 tends to be flattened, and the center M becomes relatively thick. Therefore, when the proportion of iron ore pellets increases, it is preferable to increase the ratio of coke 31 (ton / ch) charged at the center to the ore raw material 11 (ton / ch) of the first layer 10, thereby relatively increasing the amount of coke 31. As a result of the inventors' investigation into the correlation between the proportion P of iron ore pellets in the first layer 10 and the ore pile inclination angle in the first layer 10, it was concluded that by determining the above-mentioned specified value α based on the above-mentioned formula 1, taking into account the proportion P of iron ore pellets in the first layer 10, it is possible to obtain an improvement in air permeability with high accuracy.

[0051] <Reducing and dissolving process> In the reduction and melting step S3, the ore raw material 11 in the stacked first layer 10 is reduced and melted while auxiliary fuel is blown into the blast furnace by hot air blown from the tuyere 1a. The blast furnace is a continuous operation, and the reduction and melting step S3 is performed continuously. On the other hand, the stacking step S1 and the center charging step S2 are performed intermittently, and the first layer 10, the second layer 20, and the center layer 30 to be newly treated in the reduction and melting step S3 are added depending on the reduction and melting treatment status of the first layer 10 and the second layer 20 in the reduction and melting step S3.

[0052] Fig. 2 shows the state in the reduction and melting step S3. As shown in Fig. 2, a raceway A, which is a hollow portion in which coke 21 swirls and exists in a significantly sparse state, is formed near the tuyere 1a by hot air from the tuyere 1a. The temperature of this raceway A is the highest in the blast furnace 1, at about 2000°C. Adjacent to the raceway A, there is a deadman B, which is a pseudo-stagnation region of coke inside the blast furnace 1. Also, from the deadman B upward, there are a dripping zone C, a cohesive zone D, and a lumpy zone E, in that order.

[0053] The temperature in the blast furnace 1 rises from the top toward the raceway A. In other words, the temperature is highest in the lumpy zone E, followed by the cohesive zone D and the dripping zone C. For example, the lumpy zone E is between 20°C and 1200°C, whereas the deadman B is between 1200°C and 1600°C. The temperature of the deadman B varies in the radial direction, and the temperature at the center of the deadman B may be lower than that of the dripping zone C. In addition, by stably circulating hot air in the center M of the furnace, the cohesive zone D with an inverted V-shaped cross section is formed, ensuring the permeability and reducibility of the furnace.

[0054] In the blast furnace 1, the iron ore raw material 11 is first heated and reduced in the lumpy zone E. In the cohesive zone D, the ore reduced in the lumpy zone E softens and shrinks. The softened and shrunk ore descends to become dripping slag and moves to the dripping zone C. In the reduction and melting step S3, the reduction of the ore raw material 11 mainly proceeds in the lumpy zone E, and the melting of the ore raw material 11 mainly occurs in the dripping zone C. In the dripping zone C and the deadman B, direct reduction proceeds in which the descending liquid iron oxide FeO directly reacts with the carbon of the coke 21.

[0055] The aggregate 11a including the reduced iron casts exerts an aggregate effect in the cohesive zone D. In other words, even if the ore is in a state where it softens and shrinks, the reduced iron casts having a high melting point do not soften, and an air passage for reliably passing the hot air to the center of the blast furnace 1 is secured.

[0056] In addition, molten pig iron F, which is molten reduced iron, is piled up in the hearth, and molten slag G is piled up on top of the molten pig iron F. The molten pig iron F and molten slag G can be taken out from the tap hole 1b.

[0057] Examples of the auxiliary fuel injected from the tuyere 1a include pulverized coal made by pulverizing coal to a particle size of about 50 μm, heavy oil, natural gas, etc. The auxiliary fuel functions as a heat source, a reducing agent, and a recarburizer. In other words, it replaces the roles of the coke 21 except for the role of a spacer.

[0058] <Advantages> In the pig iron manufacturing method, coke 31 is charged into the center M of the blast furnace 1 for each charge in which a stacking unit consisting of the first layer 10 and the second layer 20 in the stacking step S1 is stacked, so that the flow of hot air in the center M of the blast furnace 1 is easily ensured. In addition, the permeability of the hot air is improved by setting the ratio R of the mass (ton / ch) of the coke 31 accumulated in the center M to the mass (ton / ch) of the raw ore 11 in each charge to a predetermined value α or more. Therefore, even if the amount of coke used is reduced, the necessary permeability can be ensured, so that the amount of coke can be further reduced.

[0059] [Second embodiment] Another embodiment of the method for producing pig iron according to the present invention is a method for producing pig iron using a blast furnace 1 having a tuyere 1a shown in FIG. 2, and includes a step of alternately stacking a first layer 10 containing raw ore 11 and a second layer 20 containing coke 21 in the blast furnace 1 (stacking step), a step of charging coke 31 into the center M of the blast furnace 1 (center charging step), and a step of reducing and melting the stacked raw ore 11 of the first layer 10 while blowing auxiliary fuel into the blast furnace 1 with hot air blown from the tuyere 1a (reduction and melting step). In the stacking step, the center charging step is performed one or more times during one charge in which a stacking unit consisting of one first layer 10 and one second layer 20 is stacked, and in the one charge, the ratio R of the mass (ton / ch) of the coke 31 deposited in the center M to the mass (ton / ch) of the raw ore is set to a predetermined value α or more. The pig iron manufacturing method also includes a learning process, an acquisition process, an input process, an estimation process, and a control process.

[0060] The laminating step, the central portion charging step, and the reducing and melting step are similar to the laminating step S1, the central portion charging step S2, and the reducing and melting step S3 in the first embodiment, and therefore detailed description thereof will be omitted.

[0061] (Learning process) The learning process is a process of inputting into an artificial intelligence model, as learning data, actual values ​​of an input data group including at least the temperature and blowing rate of the hot air, the furnace wall heat removal amount which is the amount of heat radiated from the furnace wall, the solution loss reaction amount which is the amount of heat due to the solution loss reaction, the amount of residual iron, the temperature of molten iron F, and the ratio R during a predetermined period from a time before a reference time to the reference time, and an output data group including temperature data of the molten iron F obtained in the reducing and melting processes in the future from the reference time, and training the artificial intelligence model to predict the temperature data of the molten iron F in the future from the reference time from the input data group.

[0062] In addition to the temperature of the hot blast and the temperature of the molten iron F, the input data set may preferably include the moisture content of the hot blast, the coke ratio, the auxiliary fuel ratio (if the auxiliary fuel contains pulverized coal, a pulverized coal ratio is preferable) and the like, from the viewpoint of improving prediction accuracy.

[0063] In the learning process, the reference time is a point in the past, and the time before the reference time may be any time as long as it is before the reference time. Also, the time in the future of the reference time is at least a time before the current time. Therefore, the numerical values ​​of the input data group and the output data group can all be actual measured values. Also, the reference time can be changed for a series of time series data from the past time to the future time, that is, the time dividing the past data and the future data can be changed.

[0064] In addition, the group of input data preferably includes data for a time later than the reference time (but earlier than the present time). The temperature of the molten iron F after the reference time, which is the subject of prediction, may change due to deliberate control of the group of input data after the reference time. Therefore, by using these data for training the artificial intelligence model, a highly accurate prediction model can be created.

[0065] The input data group and the output data group can be acquired by a sensor or the like installed in the blast furnace 1. In this case, for example, for the same type of input data, sensors may be installed at different positions and the data may be acquired as data from different locations.

[0066] The input data group and the output data group, which are actual values, are input to an artificial intelligence model as learning data. Then, the artificial intelligence model is trained to predict future temperature data of molten iron F from the reference time from the input data group. Specifically, an estimation model for predicting the temperature data of molten iron F is constructed. A known estimation technique related to machine learning (AI) can be used to construct the estimation model. Specifically, a construction means can learn the correlation between the input data group and the output data group using the input data group and the output data group, and construct an estimation model. Among them, it is preferable to use deep learning using a multi-layered neural network as the machine learning.

[0067] In the pig iron manufacturing method, the input data group acquired in the acquisition step described later and the actual values ​​of the output data group corresponding to the input data group are used as inputs in the learning step. By making predictions using a pre-trained artificial intelligence model and at the same time continuously learning using the input data group and the output data group that are successively generated by the operation of the blast furnace 1, the prediction accuracy of the artificial intelligence model can be improved.

[0068] (Acquisition process) In the acquiring step, the input data group is acquired with the current time as the reference time. Specifically, the input data group can be acquired in the same manner as the input data group used in the learning step, for example, by using the same sensor.

[0069] (Input process) In the input step, the input data group acquired in the acquisition step is input to the trained artificial intelligence model with the current time as a reference time. In the input step, the artificial intelligence model is used with the reference time as the current time, so that a time in the future from the reference time is also a future time in the real world, and the temperature of the molten iron F predicted in the estimation step described later is a temperature in the future.

[0070] (Estimated process) In the estimation step, the trained artificial intelligence model is caused to estimate the future temperature of the molten iron F. Since the artificial intelligence model is a trained model, it can accurately estimate the future temperature of the molten iron F.

[0071] (Control process) In the control step, the set values ​​of the items included in the input data group are changed based on the temperature of the molten pig iron F estimated in the estimation step. In particular, in the pig iron production method, if the ratio R is too large, the molten pig iron temperature may drop and cause cooling. For this reason, it is important to estimate the future temperature of the molten pig iron F and appropriately control it to avoid cooling.

[0072] Specifically, when a drop in the temperature of the molten iron F is predicted in the estimation step, parameters that can effectively avoid a drop in the temperature (cooling) of the molten iron F can be estimated and controlled by the artificial intelligence model from, for example, the hot air temperature, blast volume and moisture, the coke ratio, the pulverized coal ratio, and the ratio R. In particular, it is preferable to control the ratio R so that it does not become too high.

[0073] The above control process is not essential and can be omitted. In this case, the artificial intelligence model only estimates the temperature of the molten iron F. Then, the operator considers and implements specific countermeasures based on the estimation results.

[0074] <Advantages> In the molten iron production method, the temperature of the molten iron F is estimated using the trained artificial intelligence model in this way, and additional learning is performed using the input data group acquired in the acquisition step and the actual values ​​of the output data group corresponding to this input data group, making it possible to control the temperature of the molten iron F with high accuracy based on the ratio R. Therefore, it is possible to continue stable blast furnace operation.

[0075] [Other embodiments] It should be noted that the present invention is not limited to the above-described embodiment.

[0076] In the above embodiment, the pig iron manufacturing method may include other steps. For example, the pig iron manufacturing method may include a step of pulverizing the powder derived from the reduced iron molded body and coal. In this case, it is preferable to include the fine powder obtained in the pulverizing step as the auxiliary fuel. The reduced iron molded body is partially crushed into powder during the transportation process or the like. Such powder reduces the gas permeability in the blast furnace, so it is not suitable to use it as the first layer. In addition, since this powder has a large specific surface area, it is reoxidized to iron oxide. If the auxiliary fuel containing this iron oxide is injected from the tuyere, the gas permeability can be improved. Therefore, by pulverizing the powder derived from the reduced iron molded body together with coal and using the pulverized powder and the fine powder containing the coal as the auxiliary fuel to be injected from the tuyere, the reduced iron molded body can be effectively utilized and the gas permeability in the blast furnace can be improved.

[0077] Although the case where the Bell-Armor method is used as the layering process in the above embodiment has been described, other methods can also be used. As such other methods, the Bell-Less method can be mentioned. In the Bell-Less method, a rotating chute can be used to adjust the angle while layering. In this case, the central charging process may be performed continuously before and after the layering of the second layer. For example, a method of gradually increasing the amount of coke as the layering of the second layer approaches the center and continuously performing the central charging process, or a method of gradually decreasing the amount of coke following the central charging process and continuously performing the layering of the second layer, etc. may be adopted.

[0078] In the above second embodiment, the input data group acquired in the acquisition step and the performance values ​​of the output data group corresponding to this input data group are used as inputs in the learning step, but the configuration may be such that these performance values ​​are not used as inputs in the learning step. In other words, the artificial intelligence model once constructed in the learning step may continue to be used as is without additional learning.

[0079] Also, in the second embodiment, it has been described that it is good to control the ratio R in the control step so that it does not become too high. In the above control step, a method of control based on the future temperature of the molten pig iron F estimated in the estimation step has been described, but it may be based on other parameters, for example, the temperature of the molten pig iron F in real time. In other words, the pig iron manufacturing method may include a control step of controlling the ratio R so as to avoid a drop in the temperature of the molten pig iron. The above control step can suppress the adverse effect of the ratio R becoming too high. EXAMPLES

[0080] The present invention will be described in more detail below with reference to examples, but the present invention is not limited to these examples.

[0081] <Ratio R> Using operational data from a blast furnace in operation, the relationship between the ratio R of the mass of coke deposited in the center (ton / ch) to the mass of ore raw material (ton / ch) in one charge of the blast furnace and the total K value was investigated.

[0082] The total K value is the furnace top pressure P 1 (kPa), Blow pressure P 2 (kPa), Bosh gas volume BOSH (Nm 3 / min), it is expressed by the following formula 2. The lower the total K value, the more room there is for ventilation, and conversely, it enables operation with a reduced coke rate.

number

[0083] The relationship between the total K value and the lump coke ratio is Total K value = -0.0042 × lump coke ratio + 3.8054 ··· 3 Therefore, in accordance with Equation 3, the total K value was corrected to the value when the lump coke ratio was 270 kg / tp (corrected total K value), and the relationship between this corrected total K value and the ratio R was obtained. The results are shown in Figure 4. Note that "lump coke" refers to coke with a large particle size that has been sieved to ensure breathability.

[0084] From the results in Figure 4, it can be seen that the corrected total K value decreases as the ratio R increases, and that the corrected total K value is particularly low when the ratio R is greater than the specified value α = 0.017. Therefore, the relationship between the coke rate and the total K value was plotted for each operational data divided into R < 0.017 and R ≥ 0.017. The results are shown in Figure 5.

[0085] From Fig. 5, it can be seen that the total K value decreases and the permeability improves by setting R ≧ 0.017. Therefore, the coke rate can be reduced until the total K value reaches the upper limit of the operation.

[0086] <Relationship with iron ore pellet ratio P> First, the effect of the iron ore pellet ratio P on the ore pile inclination angle θ was examined.

[0087] Figure 6 shows the blast furnace burden distribution experimental apparatus 8 used in this experiment. The blast furnace burden distribution experimental apparatus 8 shown in Figure 6 is a two-dimensional slice cold model simulating a Bell-Armor type raw material charging device at a scale of 1 / 10.7. The size of the blast furnace burden distribution experimental apparatus 8 is 1450 mm in height (length of L1 in Figure 8), 580 mm in width (length of L2 in Figure 8), and 100 mm in depth (length perpendicular to the paper surface in Figure 8).

[0088] The components of the blast furnace charge distribution experimental apparatus 8 are numbered the same as the corresponding components of the Bell-Armor type raw material charging apparatus 2 in Fig. 2. Since the functions are the same, detailed explanations are omitted. In addition, as shown in Fig. 6, the blast furnace charge distribution experimental apparatus 8 has a central charging chute 8a for charging coke simulating central charging.

[0089] A base coke layer 81, a centrally charged coke layer 82, and an ore layer 83 were charged in this order into the blast furnace burden distribution experimental device 8, and then an experimental layer 84, which was an ore layer, was charged.

[0090] The raw materials used for charging the experimental layer 84 were sintered ore (particle size 2.8 to 4.0 mm) simulating sintered ore and lump ore, alumina balls (φ2 mm) simulating iron ore pellets, and coke (particle size 8.0 to 9.5 mm) simulating lump coke. The raw materials were scaled at 2 / 11.2.

[0091] Under the above conditions, the ratio of alumina balls simulating iron ore pellets was changed to measure the ore pile inclination angle θ. The target range of the ore pile inclination angle θ was set to 0.32 to 0.71 in non-dimensional radius (furnace wall side = 0.00, center side = 1.00). The results are shown in Figure 7.

[0092] From the results in Fig. 7, it was confirmed that the ore pile inclination angle θ tends to decrease by about 1 degree when the proportion of iron ore pellets (proportion of alumina balls) increases by 10%.

[0093] Next, using the operational data of a blast furnace in operation, the relationship between the coke rate and the total K value was plotted for each operational data group, R<0.017×(0.001×P+0.97) and R≧0.017×(0.001×P+0.97). The results are shown in Figure 8.

[0094] From the results in Fig. 8, it can be seen that the improvement effect on air permeability can be determined with high accuracy by taking into account the proportion P of iron ore pellets. [Industrial Applicability]

[0095] By using the pig iron manufacturing method of the present invention, it is possible to further reduce the amount of coke used while ensuring the flow of hot air in the center of the blast furnace. [Explanation of symbols]

[0096] 1 blast furnace 1a tuyere 1b Taphole 2 Raw material charging device 2a Bell Cup 2b Lower Bell 2c Armor 10 1st layer 11 Mineral ore 11a Aggregate 11b Mineral ores, excluding aggregates 12 Dripping slag 20 2nd layer 21 Coke 30 central layer 31 Coke 8. Blast Furnace Charge Distribution Experimental Device 8a Central charging chute 81 Coke layer 82 Central coke layer 83 Ore Layer 84 Experimental Layer A Raceway B Furnace core C. Droplet zone D Cohesive zone E. Lumpy Zone F Molten iron G Molten slag M Center

Claims

1. A method for producing pig iron using a blast furnace having a tuyere, comprising the steps of: stacking alternately first layers including raw ore and second layers including coke in the blast furnace; Charging coke into the center of the blast furnace; reducing and melting the ore raw material of the first layer stacked while blowing auxiliary fuel into the blast furnace with hot air blown from the tuyere; Equipped with In the laminating step, the charging step is carried out one or more times during one charge for laminating a lamination unit including one of the first layer and one of the second layer, A method for producing pig iron, in which the ratio R of the mass (ton / ch) of the coke deposited in the center to the mass (ton / ch) of the ore raw material charged in one charge is 0.017 or more.

2. A method for producing pig iron using a blast furnace having a tuyere, comprising the steps of: stacking alternately first layers including raw ore and second layers including coke in the blast furnace; Charging coke into the center of the blast furnace; reducing and melting the ore raw material of the first layer stacked while blowing auxiliary fuel into the blast furnace with hot air blown from the tuyere; Equipped with In the laminating step, the charging step is carried out one or more times during one charge for laminating a lamination unit including one of the first layer and one of the second layer, the first layer of ore feed comprises iron ore pellets; When the proportion of the iron ore pellets in the ore raw material of the first layer is P (mass%), A method for producing pig iron, in which the ratio R of the mass (ton / ch) of coke deposited in the center to the mass (ton / ch) of the ore raw material charged in one charge is equal to or greater than α, as calculated by the following formula 1. α=0.017×(0.001×P+0.97) ・・・1

3. 3. A method for producing pig iron as claimed in claim 1 or claim 2, wherein the strength of the coke deposited in the central portion is equal to or greater than the strength of the coke contained in the second layer.

4. A method for producing pig iron as described in any one of claims 1 to 3, wherein the average particle size of the coke deposited in the center portion is equal to or larger than the average particle size of the coke contained in the second layer.

5. a step of inputting into an artificial intelligence model, as learning data, actual values ​​of an input data group including at least the hot air temperature and blast volume, the solution loss reaction volume, the amount of heat removed from the furnace wall, the amount of residual iron, the molten iron temperature, and the ratio R during a predetermined period from a time that is earlier than a reference time to the reference time, and an output data group including molten iron temperature data obtained in the reducing and melting steps in the future from the reference time, and training the artificial intelligence model to predict the molten iron temperature data in the future from the reference time based on the input data group; acquiring the input data group using a current time as the reference time; inputting the input data group acquired in the acquiring step into the trained artificial intelligence model with a current time as a reference time; causing the trained artificial intelligence model to estimate a future temperature of the molten iron; Equipped with A method for producing pig iron as described in any one of claims 1 to 4, wherein the input data group acquired in the acquisition process and the actual values ​​of the output data group corresponding to this input data group are used as input for the learning process.

Images