Method for predicting blast furnace conditions, blast furnace condition prediction device, program, blast furnace operation method, and method for producing molten iron.

Abstract

A method for predicting blast furnace conditions, a blast furnace condition prediction device, a program, a blast furnace operation method, and a method for producing molten iron are provided, enabling long-term predictions and the implementation of early operational actions. [Solution] The blast furnace condition prediction method includes: a first quality prediction step in which the quality of sintered ore is predicted using a sintered ore quality prediction model that takes sintering operation performance data, which is the manufacturing condition of sintered ore produced in a sintering furnace, as an input variable and a predicted value of the quality of the sintered ore as an output variable; a second quality prediction step in which the properties of the sintered ore to be input into the blast furnace are calculated based on the predicted quality of the sintered ore; and an operational indicator prediction step in which the calculated properties of the sintered ore to be input into the blast furnace are input into a blast furnace model to predict the operational indicators of the blast furnace.

Description

Technical Field

[0001] The present disclosure relates to a method for predicting the furnace condition of a blast furnace, a device for predicting the furnace condition of a blast furnace, a program, an operation method of a blast furnace, and a method for producing hot metal.

Background Art

[0002] In the blast furnace process, ore is charged from the top of the furnace as raw material and coke as fuel to produce hot metal. In the operation method of the blast furnace, operation actions are determined based on the operation indexes of the blast furnace. For example, Patent Document 1 proposes a method for controlling a blast furnace based on prediction using a physical model. In the operation method of the blast furnace of Patent Document 1, parameters included in the physical model are adjusted from blast furnace operation data (the composition of the current top gas), a blast furnace operation prediction model and an operation action determination model are constructed to predict the operation indexes of the blast furnace, and operation actions are determined based on the prediction.

[0003] Here, sintered ore occupies most of the ore charged. The quality of the sintered ore affects the calculation of the operation indexes of the blast furnace. At present, the quality of the sintered ore used in the calculation of the operation indexes or the blast furnace operation prediction model is often a fixed value or a measured value. When a fixed value is used, due to the difference from the quality of the sintered ore actually charged into the blast furnace, the calculation of the operation indexes or the output of the blast furnace operation prediction model may deviate from the actual situation. Also, when a measured value is used, since it takes time to measure the quality of the sintered ore, there are items with a low measurement frequency (for example, several times a day). The change in quality during the measurement cannot be reflected, and the calculation of the operation indexes or the output of the blast furnace operation prediction model may deviate from the actual situation. Therefore, there is a problem that it is difficult to take an optimal operation action with respect to the change in the furnace condition (furnace situation) of the blast furnace based on the change in the quality of the sintered ore.

[0004] Regarding this problem, a method for reflecting the quality of sintered ore in the prediction of the furnace condition of a blast furnace has been proposed. For example, Patent Document 2 discloses a method of measuring the particle size and components of sintered ore in the conveying route of sintered ore to the blast furnace, calculating the properties of the raw material from the measurement results, and using them as the input to the mathematical model of the blast furnace.

Prior Art Documents

[0005] [Patent Document 1] Japanese Patent Application Publication No. 11-335710 [Patent Document 2] Japanese Patent Publication No. 2021-167447 [Overview of the project] [Problems that the invention aims to solve]

[0006] Patent Document 2 discloses a method for continuously measuring the particle size and composition of sintered ore along a transport route, calculating the raw material properties of the sintered ore, and using this information for blast furnace operation. However, because the measurement timing is immediately before charging the sintered ore into the blast furnace, it is unsuitable for long-term predictions, and early operational actions are difficult. Furthermore, there are limitations to the properties of sintered ore that can be calculated from particle size or composition, and measurement accuracy is easily degraded by disturbances in the measurement environment, making it easy for errors to occur in measured or calculated values. In addition, if there are localized biases in the shape or composition of the sintered ore during transport, the measured values ​​may not correspond to the overall properties of the sintered ore.

[0007] In view of these circumstances, the purpose of this disclosure is to provide a blast furnace condition forecasting method, a blast furnace condition forecasting device, a program, a blast furnace operation method, and a molten iron manufacturing method that enable long-term forecasting and the implementation of early operational actions. [Means for solving the problem]

[0008] (1) A method for predicting the furnace conditions of a blast furnace according to one embodiment of the present disclosure is: A first quality prediction step involves using a sintered ore quality prediction model, in which sintering operation performance data, which is the manufacturing condition of the sintered ore produced in a sintering furnace, is used as an input variable, and the predicted quality value of the sintered ore is used as an output variable, to predict the quality of the sintered ore. A second quality prediction step calculates the properties of the sintered ore to be input into the blast furnace based on the predicted quality of the sintered ore, The process includes an operational indicator prediction step of inputting the calculated properties of the sintered ore to be input into the blast furnace into a blast furnace model and predicting the operational indicators of the blast furnace.

[0009] (2) As one embodiment of the present disclosure, in (1), The second quality prediction step calculates the properties of the sintered ore to be input into the blast furnace based on the predicted quality of the sintered ore, using conveying equipment data including at least information on the belt conveyor and hopper.

[0010] (3) In one embodiment of the present disclosure, in (1) or (2), The aforementioned second quality prediction step is: The time it takes for the sintered ore to be fed into the hopper via the belt conveyor is calculated based on the speed and length of the belt conveyor. The residence time in the hopper is calculated based on the amount of sintered ore produced, the weight in the hopper, the amount extracted from the hopper, the time of extraction from the hopper, or the frequency of extraction from the hopper. The time it takes for the sintered ore extracted from the hopper to be fed into the bunker is calculated based on the speed and length of the conveyor belt. The system predicts the time at which the sintered ore is fed into the blast furnace from the top, and uses the predicted quality value of the sintered ore at the predicted time at which it is fed into the blast furnace as input to the blast furnace model at that time.

[0011] (4) In one embodiment of the present disclosure, in any of (1) to (3), The aforementioned second quality prediction step is: A dataset is obtained that includes the predicted quality and production weight of the sintered ore, categorized for each identical manufacturing condition in the sintering furnace. The aforementioned dataset is sorted in the order it is put into the hopper, The sorted dataset is grouped into extraction categories based on the unit extraction amount of sintered ore cut from the hopper, starting from the hopper outlet side. The groups of data sets divided into the aforementioned cutting sections are sorted in the order in which they are cut from the hopper onto the conveyor belt. Group the sorted data set into batch categories according to the input weight per batch to be charged into the blast furnace, For each group of the data set classified into batch categories, calculate a representative value of the quality prediction value of the sinter and use it as an input value for the blast furnace model.

[0012] (5) As one embodiment of the present disclosure, in any one of (1) to (4), The first quality prediction step predicts the quality of the sinter including at least one of particle size, strength, reducibility, reduction degradation, basicity, and FeO component.

[0013] (6) As one embodiment of the present disclosure, in any one of (1) to (5), The operation index prediction step predicts future operation indexes of the blast furnace including at least one of molten iron temperature, molten iron production rate, and air permeability index.

[0014] (7) A blast furnace condition prediction device according to an embodiment of the present disclosure, Using a sinter quality prediction model with sinter operation performance data, which is the production condition of the sinter produced in the sintering furnace, as an input variable and the quality prediction value of the sinter as an output variable, a first quality prediction unit for predicting the quality of the sinter; A second quality prediction unit for calculating the properties of the sinter input to the blast furnace based on the predicted quality of the sinter; An operation index prediction unit that inputs the calculated properties of the sinter input to the blast furnace into a blast furnace model and predicts the operation index of the blast furnace.

[0015] (8) A program according to an embodiment of the present disclosure, A computer, Using a sinter quality prediction model with sinter operation performance data, which is the production condition of the sinter produced in the sintering furnace, as an input variable and the quality prediction value of the sinter as an output variable, a first quality prediction unit for predicting the quality of the sinter; A second quality prediction unit for calculating the properties of the sinter input to the blast furnace based on the predicted quality of the sinter; Function as an operation index prediction unit that inputs the properties of the sinter ore input to the blast furnace, which have been calculated, into a blast furnace model and predicts the operation indexes of the blast furnace.

[0016] (9) The operation method of a blast furnace according to an embodiment of the present disclosure Using the furnace condition prediction method of the blast furnace according to any one of (1) to (6), calculate the properties of the sinter ore input to the blast furnace until a predetermined time in advance, input them into the blast furnace model, predict the operation indexes of the blast furnace until a predetermined time in advance, and perform the operation of the blast furnace based on the predicted operation indexes of the blast furnace.

[0017] (10) The method for producing hot metal according to an embodiment of the present disclosure Using the furnace condition prediction method of the blast furnace according to any one of (1) to (6), calculate the properties of the sinter ore input to the blast furnace until a predetermined time in advance, input them into the blast furnace model, predict the operation indexes of the blast furnace until a predetermined time in advance, and produce hot metal based on the predicted operation indexes of the blast furnace.

Advantages of the Invention

[0018] According to the present disclosure, it is possible to provide a furnace condition prediction method for a blast furnace, a furnace condition prediction device for a blast furnace, a program, an operation method for a blast furnace, and a method for producing hot metal that enable long-term prediction and implementation of early operation actions.

Brief Description of the Drawings

[0019] [Figure 1] FIG. 1 is a block diagram showing a configuration example of a furnace condition prediction device for a blast furnace according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a flowchart illustrating the process of predicting the quality of sinter ore. [Figure 3] FIG. 3 is a flowchart illustrating the process of predicting the quality of charged sinter ore. [Figure 4] FIG. 4 is a flowchart illustrating the process of predicting operation indexes. [Figure 5]Figure 5 is a flowchart illustrating the process for deciding on operational actions. [Figure 6] Figure 6 is a schematic diagram illustrating an example of a transport route for sintered ore from the sintering production line to the blast furnace. [Figure 7] Figure 7 illustrates an example of equipment that supplies the manufactured sintered ore to the blast furnace as raw material. [Figure 8] Figure 8 is a diagram illustrating the details of the quality prediction for the charged sintered ore shown in Figure 3. [Figure 9] Figure 9 is a diagram illustrating the details of the quality prediction for the charged sintered ore shown in Figure 3. [Figure 10] Figure 10 is a diagram illustrating the details of the quality prediction for the charged sintered ore shown in Figure 3. [Modes for carrying out the invention]

[0020] Hereinafter, a method for predicting the furnace conditions of a blast furnace, a blast furnace condition prediction device 10 (see Figure 1), a program, a method for operating a blast furnace, and a method for manufacturing molten iron according to one embodiment of the present disclosure will be described with reference to the drawings.

[0021] (Sintered ore transport equipment) Figure 7 shows an example of equipment for supplying manufactured sintered ore to a blast furnace as raw material. Sintered ore is an aggregate produced by sintering iron ore (powdered ore) together with limestone and coke powder in a sintering furnace to a certain size. The sintered ore produced in the sintering equipment is sent to a storage tank (or hopper) for storage. There may be multiple storage tank systems. In the example in Figure 7, there are two systems, System 1 and System 2, for two different types of ore. Each system is equipped with multiple storage tanks. In the example in Figure 7, there are two storage tanks for each system. Sintered ore is dispensed from the storage tanks in fixed amounts. The dispensed sintered ore is transported by a belt conveyor and temporarily stored in a surge hopper (or hopper). In the example in Figure 7, there is one surge hopper for each system, for a total of two surge hoppers. The surge hopper collects the ore dispensed from the hoppers for each type of ore and stores it in batches, equivalent to the amount that will be fed into the blast furnace at one time. Then, one batch (the entire amount stored in the surge hopper) of sintered ore stored in the surge hopper tank is alternately cut from the two surge hoppers onto a belt conveyor and transported to the bunker at the top of the blast furnace. In the example shown in Figure 7, three bunkers are arranged. After one batch of sintered ore has been stored in each bunker, the sintered ore is sequentially charged into the blast furnace by a bell-type or bell-less charging device.

[0022] (Device configuration) Referring to Figure 1, the configuration of the blast furnace condition prediction device 10 according to this embodiment will be explained. Figure 1 is a block diagram showing the configuration of the blast furnace condition prediction device 10. The blast furnace condition prediction device 10 is composed of an information processing device such as a computer. An internal processing unit such as a CPU (Central Processing Unit) executes a computer program to function as a first quality prediction unit 11, a second quality prediction unit 12, an operational indicator prediction unit 13, and an operational action determination unit 14. The functions of each unit will be described later.

[0023] The blast furnace condition prediction device 10 is connected to an operation database 20 in a data-readable format. In this embodiment, the operation database 20 stores blast furnace operation data such as blast furnace operation factors such as blast flow rate, blast oxygen flow rate, pulverized coal flow rate, coke ratio, blast moisture content, blast temperature, blast pressure, shaft pressure, top gas pressure, top gas temperature, and top gas components. The operation database 20 also stores sintering operation data, such as the production conditions for the sintered ore produced in the sintering furnace. The sintering operation data includes sintering operation factors such as pallet speed, exhaust gas components, exhaust gas temperature, exhaust gas flow rate, and component analysis values ​​of the sintered ore. The operation database 20 also stores transport equipment data, including information on at least the belt conveyor and hopper. The transport equipment data includes, for example, the speed of the belt conveyor on the transport route of the sintered ore from the sintering production line to the blast furnace, the amount dispensed from the hopper, and the dispensing frequency.

[0024] In the blast furnace condition prediction method executed by the blast furnace condition prediction device 10 according to this embodiment, items extracted from data stored in the operation database 20 are used. The blast furnace condition prediction device 10 comprises a first quality prediction unit 11, a second quality prediction unit 12, an operation indicator prediction unit 13, and an operation action determination unit 14, and determines the operation action of the blast furnace by executing the processes described below.

[0025] (First Quality Prediction) The operation of the first quality prediction unit 11 will be explained with reference to Figure 2. Figure 2 is a flowchart showing the process of first quality prediction (sintered ore quality prediction) performed by the first quality prediction unit 11.

[0026] In step S11, the first quality prediction unit 11 acquires sintering operation performance data from the operation database 20. The sintering operation performance data may include, for example, the blending specifications for each raw material type during sintering ore production, calcination specifications, heat source blending ratio, quicklime blending ratio, measured component values ​​of sintered ore, and the reducibility (RDI) of sintered ore. The sintering operation performance data may also include pallet speed, raw material bulk density, raw material moisture content, raw material coke content, etc. The sintering operation performance data is NO X, SO X This may include exhaust gas components such as O2, exhaust gas temperature, exhaust gas flow rate, etc.

[0027] In step S12 (first quality prediction step), the first quality prediction unit 11 predicts the quality of the sintered ore based on the sintering operation performance data acquired in step S11. In this embodiment, the first quality prediction step predicts the quality of the sintered ore using a sintered ore quality prediction model that takes sintering operation performance data as an input variable and a predicted value of the sintered ore as an output variable. The quality of the sintered ore, which affects the blast furnace operating indicators (e.g., molten iron temperature, permeability, molten iron production rate, etc.), may include, for example, particle size, strength, reducibility (RI), reducibility into pulverization (RDI), basicity, FeO component, etc., and at least one of these may be predicted. However, the quality of the sintered ore to be predicted is not limited to these. Here, if the particle size of the sintered ore is large, it becomes difficult to reduce and affects the molten iron temperature, permeability, and iron production rate of the blast furnace. The strength of the sintered ore is an indicator of its ease of collapse (pulverization) and affects the permeability of the blast furnace. The reducibility of sintered ore affects the molten iron production rate of the blast furnace and is related to the enlargement of the fusion zone within the blast furnace, thus affecting permeability. The reducibility of sintered ore affects the molten iron temperature and production rate of the blast furnace. Furthermore, as the sintered ore pulverizes and the amount of powder increases, air passages are eliminated, so the reducibility of sintered ore also affects the permeability of the blast furnace. The basicity of sintered ore affects the adjustment of the Si component or slag component of the molten iron in the blast furnace. The FeO component of sintered ore affects the molten iron temperature and production rate of the blast furnace.

[0028] Here, the quality of the sintered ore may be predicted by known methods. For example, the quality of the sintered ore may be predicted using the statistical model described in Reference 1 (Japanese Patent Publication No. 2024-015703) as a sintered ore quality prediction model. The statistical model in Reference 1 takes as input variables the blending specifications, calcination specifications, heat source blending ratio, quicklime blending ratio, and sintered ore component values ​​for each raw material type during sintered ore production, as well as sintered ore operational performance data, and the quality value of the sintered ore as the output variable. In the method of Reference 1, Lasso regression is performed to select sintered ore operational performance data that have a high contribution to the quality value of the sintered ore. Then, the quality of the sintered ore is predicted using a statistical model (Ridge regression, multiple regression analysis, support vector machine, decision tree, random forest, etc.) with the selected sintered ore operational performance data as the input variable and the sintered ore quality value as the output variable.

[0029] Furthermore, the proportion of FeO in the sintered ore may be predicted using a machine learning model. The machine learning model may be a linear model such as multiple regression analysis, a hierarchical model, a neural network, GBDT, a random forest, a decision tree, or a transformer. The machine learning model may predict the proportion of FeO in the exhaust gas (NO). X The main input variables may be the raw material brand, raw material moisture content, water flow rate in the granulation mixer, mixing ratio of quicklime, ore, and return ore, and exhaust gas components (CO, SO). X The following may further include CO2, cooler air pressure, cooler exhaust gas temperature, pallet speed, sintered ore layer thickness, or production volume. Hereinafter, the quality value of the sintered ore predicted by the first quality prediction unit 11 will be referred to as the "sintered ore quality prediction value".

[0030] (Second Quality Prediction) The operation of the second quality prediction unit 12 will be explained with reference to Figure 3. Figure 3 is a flowchart of the second quality prediction (quality prediction of charged sintered ore) performed by the second quality prediction unit 12.

[0031] In step S21, the second quality prediction unit 12 acquires sintering operation performance data, transport equipment data, and sintered ore quality prediction values. Sintering operation performance data and transport equipment data are acquired from the operation database 20. Sintered ore quality prediction values ​​are acquired from the first quality prediction unit 11. Transport equipment refers to equipment in the transport route of sintered ore from the sintering production line to the blast furnace. Transport equipment includes a hopper for temporarily storing sintered ore, a storage tank, a belt conveyor for transporting sintered ore cut from the hopper, and a bunker located at the top of the blast furnace for temporarily storing sintered ore. Transport equipment data includes, for example, the weight in the hopper, the weight in the bunker, the amount cut from the hopper and bunker, the cutting speed, the cutting time and cutting frequency, and the speed and length of the belt conveyor. Sintered ore quality prediction values ​​are, as described above, for example, the particle size, strength, reducibility (RI), reducibility into powder (RDI), basicity, FeO component, etc.

[0032] In step S22 (second quality prediction step), the second quality prediction unit 12 predicts the quality of the charged sintered ore (sintered ore charged from the top of the furnace) based on the transport equipment data and sintered ore quality prediction value acquired in step S21. In this embodiment, the second quality prediction step calculates the properties of the sintered ore input to the blast furnace (i.e., charged sintered ore) based on the predicted quality of the sintered ore (quality prediction value). Figure 6 shows an example of a transport route for the sintered ore. One embodiment of the method for predicting the quality of the charged sintered ore may be the method described below. For example, the sintering furnace is designated as the "sintered ore quality prediction point," and a dataset is created consisting of data linked to the time at the sintered ore quality prediction point. The dataset includes production volume and quality prediction value. The time and production volume are obtained from sintering operation performance data. The unit of production volume is, for example, [t / min]. However, manufacturing weight may be used as the production volume. The quality prediction value is the sintered ore quality prediction value acquired in step S21. Here, as shown in the table in Figure 6, a dataset is created by linking time with the production volume and predicted sintered ore quality at that time. Here, the production volume and predicted sintered ore quality at each time may be the average of the production volume and the predicted sintered ore quality over the most recent fixed period, for example, 30 minutes or 1 hour. First, the produced sintered ore is transported from the sintering production line to the hopper by a belt conveyor. Based on the speed and length of the belt conveyor included in the transport equipment data, the time it takes for the sintered ore to be fed into the hopper by the belt conveyor is calculated (see ΔT1 in Figure 6). The residence time in the hopper can be calculated based on the sintered ore production volume and the weight in the hopper, hopper dispensing volume, hopper dispensing time, or hopper dispensing frequency included in the transport equipment data (see ΔT1 in Figure 6). HP(See reference). Here, the weight inside the hopper can be determined from the initial weight inside the hopper, the amount of sintered ore produced, and the amount excavated from the hopper. The time until the sintered ore is fed into the bunker can be estimated by a similar calculation for the belt conveyor from the hopper to the bunker (see ΔT2 in Figure 6). For example, the time until the sintered ore excavated in the hopper is fed into the bunker is similarly calculated based on the speed and length of the belt conveyor. By estimating when the sintered ore produced in the sintering production line is charged from the top of the furnace (how much time has passed before it is charged), it becomes possible to use a quality prediction value that takes into account the time required for transport as input to the blast furnace model. In other words, the time at which the sintered ore produced in the sintering furnace is fed from the top of the blast furnace is predicted, and the predicted quality prediction value of the sintered ore at the predicted top-of-furnace time is used as input to the blast furnace model at the top-of-furnace time. Here, the transport route for the sintered ore is not limited to the configuration in Figure 6; for example, there may be multiple transport routes in parallel, and there may be multiple hoppers for a single transport route. Here, the quality value of the charged sintered ore predicted by the second quality prediction unit 12 is referred to as the "charged sintered ore quality prediction value."

[0033] (Methods to reduce load) As described above, the quality prediction of charged sintered ore reflects the process from the sintered ore quality prediction point (sintering furnace) to its charging into the blast furnace. Here, after being produced in the sintering furnace, the sintered ore is temporarily stored in multiple storage tanks (raw material tanks). Subsequently, when it is used as raw material in the blast furnace, in order to adjust the supply amount, the sintered ore is temporarily stored in a hopper (surge hopper), and a predetermined amount (equivalent to one batch) is drawn out. The sintered ore drawn out from the hopper is stored in a bunker located at the top of the furnace via a belt conveyor, and then charged into the blast furnace by a bell-type or bell-less charging device. In this way, the sintered ore produced in the sintering furnace passes through multiple temporary storage tanks such as storage tanks, hoppers, and bunkers, and a predetermined amount is drawn out from each temporary storage tank. Therefore, it is rare for sintered ore produced in one batch in the sintering furnace (i.e., sintered ore produced under the same production conditions) to become raw material for one batch. In other words, in most cases, sintered ore of different qualities are mixed in each batch. Therefore, while it is possible to perform precise calculations based on the history of sintered ore, this generally results in a heavy computational load.

[0034] Therefore, the following describes a method to reduce the computational load by handling data in a simplified manner, while reflecting the process up to the point of charging the blast furnace. To enable this simplified data handling, it is assumed that the conditions described in the following paragraphs are met.

[0035] The quality of sintered ore produced in a sintering furnace is determined on a per-production basis, or on a per-production basis under the same production conditions. Furthermore, the quality of sintered ore produced in a sintering furnace is determined by a dataset (sintered ore dataset, see table in Figure 6) that includes time (actual production time), production volume (production weight), and quality prediction as basic components. When the sintered ore produced in the sintering furnace is transported by a belt conveyor, it is sorted according to production conditions and loaded onto the belt conveyor. In other words, it is not mixed on the belt conveyor. The sintered ore loaded onto the belt conveyor is fed into a storage tank, hopper, or bunker in the order it was loaded. If the sintered ore is not agitated within the tank, it is dispensed from the discharge port in predetermined dispensing units in the order it was fed. If agitation is performed within the tank, the ore is homogenized (averaged) within the tank before being dispensed in predetermined dispensing units. If the sintered ore placed in the bunker is not stirred, it is fed into the blast furnace in batches (one batch's worth) in the order it was placed in the bunker. If the sintered ore placed in the bunker is stirred, it becomes homogenized (averaged) within the bunker and is then fed into the blast furnace in batches. Calculations using the blast furnace model are performed at predetermined calculation cycles (e.g., every 30 minutes).

[0036] The process for inputting an appropriate sintered ore dataset into the blast furnace model is explained, reflecting the process from the quality prediction point (sintering furnace) of the sintered ore to its charging into the blast furnace as raw material. The second quality prediction unit 12 performs the processes from (S1) to (S6). Figure 8 is an explanatory diagram of the processes from (S1) to (S4). Figure 9 is an explanatory diagram of the process in (S5). Figure 10 is an explanatory diagram of (S6). In Figures 8 and 9, only the first system from Figure 7 is extracted and shown for the sake of clarity.

[0037] In process (S1), a sintered ore dataset is obtained. Sintered ore datasets are obtained that are separated by the same manufacturing conditions in the sintering furnace (S1 in Figure 8).

[0038] In process (S2), the acquired sintered ore dataset is sorted in the order it will be fed into the hopper (or storage tank). Based on the belt conveyor speed and length, the time it takes for the sintered ore to be fed from the sintering furnace to the hopper is calculated, and the scheduled feeding time after transport by the belt conveyor is calculated from the actual production time of the sintered ore in the sintering furnace. Then, based on the scheduled feeding time, the sintered ore dataset is rearranged in the order in which it will be fed into the hopper. If there are multiple hoppers, the sintered ore dataset is sorted in the order in which it will be fed into each hopper (S2 in Figure 8).

[0039] In process (S3), the sintered ore dataset is divided into groups (extraction sections) based on the amount extracted from the hopper (or storage tank). The sintered ore dataset is divided by the unit extraction amount, starting from the hopper outlet side (extraction side). The unit extraction amount is a predetermined weight in this embodiment. The manufacturing weight of the sintered ore dataset is accumulated starting from the hopper outlet side. The accumulated weight is divided by the unit extraction amount from the outlet side, and the sintered ore dataset at the boundary of the extraction section is divided into two sintered ore datasets (S3 in Figure 8).

[0040] In process (S4), the sintered ore datasets, which have been divided into groups, are sorted in the order in which they will be cut from the hopper onto the conveyor belt. Here, if there is one hopper, the order of the sintered ore datasets is the same as the order within the hopper. If there are multiple hoppers, the sintered ore datasets are rearranged on the conveyor belt in the order according to the pre-set cutting order of the hoppers (S4 in Figure 8).

[0041] In process (S5), the sintered ore dataset is divided into groups (batch divisions) based on the weight of one batch of material charged into the blast furnace. The sintered ore cut from the hopper onto the conveyor belt is stored in the surge hopper one batch at a time. The sintered ore dataset after process (S4) is divided into batch divisions based on the order in which it is fed into the surge hopper (S5 in Figure 9). Sintered ore datasets at the boundary of batch divisions are split into two sintered ore datasets at the boundary.

[0042] In process (S6), input values ​​for the blast furnace model are calculated based on the sintered ore datasets divided into batches. As shown in Figure 10, one batch of sintered ore stored in the surge hopper tank is alternately cut from the first and second surge hoppers onto a conveyor belt and fed into the bunkers (A, B, C) at the top of the blast furnace, one batch at a time. As input values ​​for the blast furnace model, representative values ​​of the quality prediction for each sintered ore dataset, grouped into batches, are calculated for each bunker. The representative values ​​of the quality prediction may be, for example, a weighted average of the quality predictions based on the manufacturing weight of each sintered ore dataset included in the batch, a simple average (arithmetic mean) of the quality predictions, a moving average, or the median.

[0043] As described above, sintered ore datasets are obtained, separated by identical manufacturing conditions in the sintering furnace (S1). The sintered ore datasets are sorted in the order they are fed into the hopper (S2). The sorted sintered ore datasets are grouped into extraction categories based on the unit amount of sintered ore extracted from the hopper, starting from the hopper outlet side (S3). The groups of sintered ore datasets separated by extraction categories are sorted in the order they are extracted from the hopper onto the conveyor belt (S4). The sorted sintered ore datasets are grouped into batch categories based on the input weight per batch fed into the blast furnace (S5). For each group of sintered ore datasets separated by batch categories, a representative value of the predicted quality of the sintered ore is calculated and used as an input value for the blast furnace model (S6). The processes from (S1) to (S6) reflect the process from when the sintered ore produced in the sintering furnace is charged into the blast furnace as raw material. Furthermore, by simply rearranging the order of the sintered ore dataset, it is possible to calculate a representative value for the predicted quality of the sintered ore for each batch fed into the blast furnace.

[0044] (Forecast of operational indicators) The operation of the operational indicator prediction unit 13 will be explained with reference to Figure 4. Figure 4 is a flowchart showing the process of predicting the blast furnace's operational indicators performed by the operational indicator prediction unit 13.

[0045] In step S31, the operational indicator prediction unit 13 acquires blast furnace operation performance data and predicted values ​​for charged sintered ore quality. The blast furnace operation performance data is obtained from the operation database 20.

[0046] In step S32, the operational indicator prediction unit 13 creates an input dataset for the blast furnace model based on the acquired blast furnace operation performance data and the predicted quality values ​​of the charged sintered ore. The predicted quality values ​​of the charged sintered ore include, for example, the particle size, strength, reducibility (RI), reducibility into pulverizable material (RDI), basicity, and FeO component of the sintered ore from the present to 8 hours in the future. The blast furnace operation performance data includes, for example, the current blast flow rate, blast oxygen flow rate, pulverized coal flow rate, coke ratio, blast moisture content, blast temperature, and top gas components.

[0047] In step S33 (operation indicator prediction step), the operation indicator prediction unit 13 calculates (predicts) the blast furnace operation indicator based on the input dataset it has created. In this embodiment, the operation indicator prediction step inputs the calculated properties of the sintered ore to be input into the blast furnace into the blast furnace model and predicts the blast furnace operation indicator. The operation indicator prediction unit 13 calculates the blast furnace operation indicator assuming that the current blast furnace operation performance data will be retained (unchanged) for, for example, 10 hours in advance. Here, the prediction of the blast furnace operation indicator may be performed using a physical model (unsteady-state model) that can calculate the internal state of the blast furnace (furnace) in an unsteady state as the blast furnace model. The physical model may consist of a group of partial differential equations that take into account physical phenomena such as ore reduction, heat exchange between ore and coke, and ore melting. For example, the physical model can be the model used in the method described in Reference 2 (K. Takatani et al. ISIJ International, Vol.39 (1999), pp.15). Furthermore, the physical model can be one used in the method described in Reference 3 (Michiharu Hatano et al.: "Study of Firing Operation using a Blast Furnace Unsteady-State Model", Iron and Steel, vol. 68, p. 2369). The input variables of the physical model may include, in addition to the blast flow rate, blast oxygen flow rate, pulverized coal flow rate, coke ratio, blast moisture content, blast temperature, and furnace top gas pressure, raw material properties such as the reducibility of the ore, slag ratio, coke moisture content, and particle size of coke and ore. In this embodiment, at least one of the following is used as the raw material properties, which are input variables of the physical model: particle size of sintered ore, strength, reducibility (RI), reducibility into pulverization (RDI), basicity, and FeO component.

[0048] Examples of blast furnace operating indicators include furnace heat indicators such as molten iron temperature, molten iron production volume, molten iron production rate, or permeability indicators. For example, future blast furnace operating indicators including at least one of molten iron temperature, molten iron production rate, and permeability indicators may be predicted. In addition, examples of operating indicators output from a physical model include gas utilization rate, solution loss carbon amount, or reducing agent ratio. However, the blast furnace operating indicators to be predicted are not limited to these. Hereinafter, the blast furnace operating indicators predicted by the operating indicator prediction unit 13 will be referred to as "predicted operating indicator values."

[0049] (Decision on operational actions) The operation of the operation action determination unit 14 will be explained with reference to Figure 5. Figure 5 is a flowchart showing the process of determining the operation action of the blast furnace performed by the operation action determination unit 14.

[0050] In step S41, the operation action determination unit 14 acquires blast furnace operation performance data and predicted values ​​for operation indicators.

[0051] In step S42, the operational action determination unit 14 determines the blast furnace operational action so that the blast furnace operational indicators converge to control values ​​quickly when it is predicted that the furnace conditions will deteriorate from the present onward. The furnace conditions from the present onward are predicted based on blast furnace operational performance data and predicted operational indicator values. The blast furnace operational action is, for example, an increase or decrease in operational variables such as coke ratio, pulverized coal flow rate, blown air moisture content, blown air flow rate, and blown air oxygen flow rate. However, the blast furnace operational action is not limited to these. For example, known methods may be applied to control the blast furnace's molten iron temperature, ironmaking rate, and aeration rate so that the molten iron temperature reaches a target value (target molten iron temperature), the ironmaking rate reaches a target value (target ironmaking rate), and the aeration rate reaches a control value (upper limit of furnace pressure drop). Here, for example, the process control method described in Reference 4 (International Publication No. 2024 / 048310) may be used. Furthermore, based on the acquired observed or calculated values, the molten iron temperature, ironmaking rate, and aeration rate may be controlled simultaneously. At this time, one or more operational actions from among the following may be determined: blown air flow rate, blown oxygen flow rate, pulverized coal ratio, blown air moisture content, blown air temperature, coke ratio, and furnace top pressure.

[0052] The blast furnace condition prediction method can be used in blast furnace operation methods and molten iron manufacturing methods. For example, using the above blast furnace condition prediction method, the properties of the sintered ore to be input into the blast furnace can be calculated up to a predetermined time in advance and input into a blast furnace model, and the blast furnace operation indicators can be predicted up to a predetermined time in advance. Then, a blast furnace operation method can be performed in which the blast furnace is operated based on the predicted blast furnace operation indicators, and a molten iron manufacturing method can be performed in which molten iron is produced based on the predicted blast furnace operation indicators.

[0053] As described above, the blast furnace condition prediction method, blast furnace condition prediction device 10, program, blast furnace operation method, and molten iron manufacturing method according to this embodiment enable long-term prediction and early implementation of operational actions through the above configuration. In the method disclosed herein, the predicted quality values ​​of sintered ore in the sintered ore production line, calculated using a sintered ore quality prediction model, are used as input to the blast furnace model by predicting the timing of sintered ore charging, thereby enabling long-term prediction of the blast furnace condition that reflects the quality of the sintered ore. Furthermore, based on the long-term prediction of the blast furnace condition, appropriate operational actions can be implemented earlier compared to conventional technologies. In addition, the predicted quality values ​​of sintered ore calculated using a sintered ore quality prediction model are used as input to the blast furnace model instead of measured values. Therefore, while there are limitations to the properties of sintered ore that can be calculated from particle size or composition obtained as measured values, it is possible to predict a variety of qualities. Furthermore, even if there is a localized bias in the sintered ore during transport, a representative value of the predicted quality of the sintered ore can be used, making it possible to predict the overall quality.

[0054] While embodiments of this disclosure have been described based on the drawings and examples, it should be noted that those skilled in the art will find it easy to make various modifications or alterations based on this disclosure. Therefore, it should be noted that these modifications or alterations are included within the scope of this disclosure. For example, the functions included in each component or process can be rearranged in a logically consistent manner, and multiple components or processes can be combined into one or divided. Embodiments relating to this disclosure can also be realized as storage media recording programs executed by a processor in the device. It should be understood that these are also included within the scope of this disclosure. [Explanation of Symbols]

[0055] 10. Blast furnace condition prediction device 11. First Quality Prediction Department 12. Second Quality Forecasting Department 13. Operational Indicators Forecasting Department 14. Operational Action Decision Department 20. Operations Database

Claims

1. A first quality prediction step involves predicting the quality of the sintered ore using a sintered ore quality prediction model, in which sintering operation performance data, which is the manufacturing condition of the sintered ore produced in a sintering furnace, is used as the input variable, and the predicted quality value of the sintered ore is used as the output variable. A second quality prediction step calculates the properties of the sintered ore to be input into the blast furnace based on the predicted quality of the sintered ore, A method for predicting the furnace conditions of a blast furnace, comprising: an operational indicator prediction step of inputting the calculated properties of the sintered ore to be input into the blast furnace into a blast furnace model and predicting the operational indicators of the blast furnace.

2. The blast furnace condition prediction method according to claim 1, wherein the second quality prediction step calculates the properties of the sintered ore to be input to the blast furnace based on the predicted quality of the sintered ore, using conveying equipment data including information on at least a belt conveyor and a hopper.

3. The aforementioned second quality prediction step is: The time it takes for the sintered ore to be fed into the hopper via the belt conveyor is calculated based on the speed and length of the belt conveyor. The residence time in the hopper is calculated based on the amount of sintered ore produced, the weight in the hopper, the amount extracted from the hopper, the time of extraction from the hopper, or the frequency of extraction from the hopper. The time it takes for the sintered ore extracted from the hopper to be fed into the bunker is calculated based on the speed and length of the conveyor belt. A method for predicting the furnace conditions of a blast furnace according to claim 1 or 2, wherein the time at which the sintered ore is fed into the top of the blast furnace is predicted, and the predicted quality prediction value of the sintered ore at the predicted time at the top of the furnace is used as the input to the blast furnace model at the time at the top of the furnace.

4. The aforementioned second quality prediction step is: A dataset is obtained that includes the predicted quality and production weight of the sintered ore, categorized for each identical manufacturing condition in the sintering furnace. The aforementioned dataset is sorted in the order it is put into the hopper, The sorted dataset is grouped into extraction categories based on the unit extraction amount of sintered ore cut from the hopper, starting from the hopper outlet side. The groups of data sets divided into the aforementioned cutting sections are sorted in the order in which they are cut from the hopper onto the conveyor belt. The sorted dataset is then grouped into batches based on the weight of each batch fed into the blast furnace. A method for predicting the furnace conditions of a blast furnace according to claim 1 or 2, wherein a representative value of the predicted quality of sintered ore is calculated for each group of the dataset divided into batch categories, and this is used as an input value for the blast furnace model.

5. The first quality prediction step is to predict the quality of a sintered ore, which includes at least one of particle size, strength, reducibility, reducibility into powder, basicity, and FeO component, as described in claim 1 or 2.

6. The blast furnace condition prediction method according to claim 1 or 2, wherein the operational indicator prediction step predicts future operational indicators of the blast furnace, which include at least one of molten iron temperature, molten iron production rate, and permeability index.

7. A first quality prediction unit predicts the quality of sintered ore using a sintered ore quality prediction model, which takes sintering operation performance data, which is the manufacturing condition of sintered ore produced in a sintering furnace, as an input variable and a predicted quality value of the sintered ore as an output variable. A second quality prediction unit calculates the properties of the sintered ore to be input into the blast furnace based on the predicted quality of the sintered ore, A blast furnace condition prediction device comprising: an operational indicator prediction unit that inputs the calculated properties of the sintered ore to be input into the blast furnace into a blast furnace model and predicts the operational indicators of the blast furnace.

8. Computers, A first quality prediction unit predicts the quality of sintered ore using a sintered ore quality prediction model, which takes sintering operation performance data, which is the manufacturing condition of sintered ore produced in a sintering furnace, as an input variable and a predicted quality value of the sintered ore as an output variable. A second quality prediction unit calculates the properties of the sintered ore to be input into the blast furnace based on the predicted quality of the sintered ore, A program that functions as an operational indicator prediction unit, which inputs the calculated properties of the sintered ore to be input into the blast furnace into a blast furnace model and predicts the operational indicators of the blast furnace.

9. A method for operating a blast furnace, comprising: using the blast furnace condition prediction method according to claim 1 or 2, calculating the properties of sintered ore to be input into the blast furnace up to a predetermined time in advance and inputting them into the blast furnace model; predicting the blast furnace operating indicators up to a predetermined time in advance; and operating the blast furnace based on the predicted blast furnace operating indicators.

10. A method for manufacturing molten iron, comprising: using the blast furnace condition prediction method for a blast furnace according to claim 1 or 2, calculating the properties of sintered ore to be input into the blast furnace up to a predetermined time in advance and inputting them into the blast furnace model; predicting the blast furnace operating indicators up to a predetermined time in advance; and manufacturing molten iron based on the predicted blast furnace operating indicators.

Images