Si Concentration Prediction Method, Operation Guidance Method, Si Concentration Prediction Device, Operation Guidance Device, Operation Guidance System, and Terminal Device

Abstract

The silicon concentration prediction method includes: a prediction step (S1) for predicting a future prediction value and a current value of silicon concentration using a statistical model and physical model in which an operating variable of a blast furnace is used as input and silicon concentration is used as output; and an adjustment amount calculation step (S3) for, if the predicted value or the current value largely deviate from a target value, calculating the adjustment amount of the charging amount of a desiliconization material for the operation action that reduces the silicon concentration by charging the desiliconization material, or for calculating the adjustment amount of the operating variable for the operation action that controls the heat level in the blast furnace in order to increase the silicon concentration at the time of tapping before desiliconization.

Description

Technical Field

[0001] The present disclosure relates to a Si concentration prediction method, an operation guidance method, a Si concentration prediction device, an operation guidance device, an operation guidance system, and a terminal device.

Background Art

[0002] In the blast furnace method and the converter method in the iron and steel industry, the Si concentration in hot metal (hot metal Si concentration) is an important variable because it is an index of products for the blast furnace and a component of raw materials for the converter. Under the recent social demand for CO2 reduction, the reduction of the hot metal charge ratio (charge ratio) in the converter has been promoted. Therefore, Si in hot metal is actively utilized as a heat source in the converter.

[0003] When the Si concentration in hot metal is too low, it is necessary to input surplus heat sources (such as FeSi) in the converter, leading to an increase in raw material costs. On the other hand, when the Si concentration in hot metal is too high, slopping may occur in the pretreatment process, resulting in a decrease in productivity, or it may be necessary to input surplus quicklime for basicity adjustment. Since there is an optimal point for the Si concentration, it is necessary for the blast furnace side to adjust the Si concentration according to the optimal point.

[0004] Various proposals have been made for the prediction technology of Si concentration in the blast furnace. For example, Patent Document 1 proposes a method of calculating a predicted value of the future hot metal temperature when the operation amount is maintained, constructing a regression equation based on the calculated predicted value, and predicting the future Si concentration of hot metal by inputting it into the regression equation.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] Here, in order to adjust the Si concentration, operation actions such as reducing the Si concentration by adding a desiliconizing agent on the casting floor or changing the Si concentration at the time of tapping before desiliconization by controlling the heat level in the blast furnace can be considered. Here, the heat level is a section obtained by dividing the temperature in the furnace by a temperature range. The temperature range may be constant or may vary. In the current blast furnace operation, since it takes time to analyze the Si concentration, the actual value can be obtained 1 to 2 hours after tapping. Therefore, when implementing the action of adding a desiliconizing agent on the casting floor, it is necessary to accurately estimate (predict) the current Si concentration. In addition, when implementing the action of changing the Si concentration by controlling the heat level in the blast furnace, since the influence of the thermal inertia of the blast furnace process is large and the immediate effect of the operation action is weak, prediction more than 2 hours ahead is essential. Therefore, a high-precision prediction method that can cope with either of these operation actions is required.

[0007] In addition, cooperation between the processes of steelmaking (from the blast furnace side) and steelmaking is required. For example, if the operator of steelmaking can appropriately communicate to the operator of steelmaking whether to add a desiliconizing agent after tapping according to the required heat source amount in the blowing charge of the converter on the steelmaking side, it is considered that it will lead to further reduction of steelmaking costs. However, the conventional method has been limited to predicting the Si concentration in the blast furnace process alone.

[0008] An object of the present disclosure made to solve the above problems is to provide a Si concentration prediction method, an operation guidance method, a Si concentration prediction device, an operation guidance device, an operation guidance system, and a terminal device capable of predicting the Si concentration with high accuracy.

Means for Solving the Problems

[0009] (1) The Si concentration prediction method according to an embodiment of the present disclosure is a prediction step of predicting a future predicted value and a current value of the Si concentration using a statistical model and a physical model that take operation variables of a blast furnace as an input and the Si concentration as an output, When the predicted value or the current value deviates significantly from the target value, an adjustment amount calculation step for calculating an adjustment amount of the input amount of the desiliconizing material in an operation action of reducing the Si concentration by inputting the desiliconizing material or an adjustment amount of the operation variable in an operation action of controlling the heat level in the blast furnace to increase the Si concentration at the time of tapping before desiliconization is included.

[0010] (2) As one embodiment of the present disclosure, in (1), In the prediction step, the statistical model is used for predicting the current value and the predicted value up to a predetermined time, and the physical model is used for predicting the predicted value after the predetermined time. The predetermined time is determined based on the time when the prediction accuracy of the statistical model decreases compared to the prediction accuracy of the physical model.

[0011] (3) As one embodiment of the present disclosure, in (1) or (2), The operation variables that are inputs to the statistical model include at least one of the tuyere injection temperature, the furnace heat index, the solution loss carbon amount, the top gas temperature, the tuyere tip gas temperature, and the specific blast volume.

[0012] (4) As one embodiment of the present disclosure, in any one of (1) to (3), The operation variables that are inputs to the physical model include at least one of the blast flow rate, the blast oxygen flow rate, the pulverized coal flow rate, the coke ratio, the blast moisture, and the blast temperature.

[0013] (5) The operation guidance method according to one embodiment of the present disclosure includes a step of presenting the predicted value and the current value predicted by any of the Si concentration prediction methods in (1) to (4) to an operator.

[0014] (6) The operation guidance method according to one embodiment of the present disclosure includes an input step in which a corrected target value of the Si concentration is input. The adjustment amount calculation step calculates the adjustment amount of the operation variable using the corrected target value input in the input step as the target value.

[0015] (7) As one embodiment of the present disclosure, in (6), The corrected target value is set based on input information from an operator other than the milling operator.

[0016] (8) The Si concentration prediction device according to one embodiment of the present disclosure includes a storage unit that stores a statistical model and a physical model that take the operation variables of the blast furnace as input and output the Si concentration, a prediction unit that predicts the future predicted value and the current value of the Si concentration using the statistical model and the physical model, and an adjustment amount calculation unit that calculates the adjustment amount of the input amount of the desiliconizing material in the operation action of reducing the Si concentration by inputting the desiliconizing material or the adjustment amount of the operation variable in the operation action of controlling the heat level in the blast furnace to increase the Si concentration at the time of tapping before desiliconization when the predicted value or the current value deviates significantly from the target value.

[0017] (9) The operation guidance device according to one embodiment of the present disclosure includes a storage unit that stores a statistical model and a physical model that take the operation variables of the blast furnace as input and output the Si concentration, a prediction unit that predicts the future predicted value and the current value of the Si concentration using the statistical model and the physical model and presents the predicted value and the current value to the operator, and an adjustment amount calculation unit that calculates the adjustment amount of the input amount of the desiliconizing material in the operation action of reducing the Si concentration by inputting the desiliconizing material or the adjustment amount of the operation variable in the operation action of controlling the heat level in the blast furnace to increase the Si concentration at the time of tapping before desiliconization when the predicted value or the current value deviates significantly from the target value.

[0018] (10) The operation guidance system according to one embodiment of the present disclosure includes an operation guidance device and a terminal device, The operation guidance device or the terminal device a storage unit that stores a statistical model and a physical model that take the operation variables of the blast furnace as inputs and output the Si concentration; a prediction unit that predicts future predicted values and current values of the Si concentration using the statistical model and the physical model, and outputs the predicted values and the current values; when the predicted value or the current value deviates significantly from the target value, calculates an adjustment amount of the input amount of the desiliconizing material in the operation action of reducing the Si concentration by inputting the desiliconizing material, or an adjustment amount of the operation variable in the operation action of controlling the heat level in the blast furnace to increase the Si concentration at the time of tapping before desiliconization, and outputs the adjustment amount; an acquisition unit that acquires the predicted value, the current value, and the adjustment amount; an input unit into which a corrected target value of the Si concentration and a future time or time zone to be corrected are input; a transmission unit that transmits the input corrected target value and the future time or time zone to be corrected; a display unit that displays the acquired predicted value, current value, adjustment amount, corrected target value, and future time or time zone to be corrected.

[0019] (11) The terminal device according to an embodiment of the present disclosure Using the future predicted value and the current value of the Si concentration predicted by a statistical model and a physical model that take the operation variables of the blast furnace as inputs and output the Si concentration, when the predicted value or the current value deviates significantly from the target value, the adjustment amount of the input amount of the desiliconizing material in the operation action of reducing the Si concentration by inputting the desiliconizing material, or the adjustment amount of the operation variable in the operation action of controlling the heat level in the blast furnace to increase the Si concentration at the time of tapping before desiliconization is calculated, an acquisition unit that acquires the predicted value, the current value, and the adjustment amount; an input unit into which a corrected target value of the Si concentration and a future time or time zone to be corrected are input; a transmission unit that transmits the input corrected target value and the future time or time zone to be corrected; A display unit that displays the obtained predicted value, the current value, the adjustment amount, the corrected target value, and the future time or time zone to be corrected.

[0020] (12) As one embodiment of the present disclosure, in (11), It includes an inter-terminal communication unit that transmits and receives information regarding the corrected target value to and from another terminal device.

Advantages of the Invention

[0021] According to the present disclosure, it is possible to provide a Si concentration prediction method, an operation guidance method, a Si concentration prediction device, an operation guidance device, an operation guidance system, and a terminal device capable of predicting the Si concentration with high accuracy.

Brief Description of the Drawings

[0022]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Modes for Carrying Out the Invention

[0023] Hereinafter, a Si concentration prediction method, an operation guidance method, a Si concentration prediction device, an operation guidance device, an operation guidance system, and a terminal device according to an embodiment of the present disclosure will be described with reference to the drawings. In the present disclosure, a statistical model and a physical model are used. The statistical model and the physical model are selectively used according to the prediction time, and highly accurate predictions are combined and presented to, for example, an operator. Hereinafter, the determination mechanism of the Si concentration, the statistical model, and the physical model will be described in order. In the description of the present embodiment, the Si concentration means the "Si concentration in hot metal", that is, the hot metal Si concentration.

[0024] The following two mechanisms are known for the determination mechanism of the Si concentration in the blast furnace. Here, [x] in the reaction formula indicates the component x in hot metal, and (x) means the component x in slag. First, as the first mechanism, SiO gas is generated in the combustion part of the tuyere of the blast furnace (SiO2 + C → SiO + CO) and absorbed by the hot metal dripping in the lower part of the furnace (SiO + [C] → [Si] + CO). As the second mechanism, Si is taken into the hot metal through the SiO2 reduction reaction ((SiO2) + 2C → [Si] + 2CO) at the slag-metal interface at the bottom of the blast furnace.

[0025] The thermodynamic equilibrium of [Si] in hot metal exceeds 10%. However, the actually measured Si concentration is at most about 2%. Therefore, it is considered that [Si] has not reached equilibrium. Also, in both mechanisms, the thermal state in the furnace has a great influence. Therefore, in the present embodiment, factors related to furnace heat are considered in the calculation of the predicted value of the Si concentration.

[0026] The statistical model is not limited to a specific model. For example, an LW-PLS (Locally-Weighted Partial Least Squares) model can be adopted as in the present embodiment. In the LW-PLS model, it is possible to predict the Si concentration while emphasizing past data similar to the current operating conditions and suppressing overfitting caused by multicollinearity.

[0027] As shown in Fig. 1, among the input variables given to the statistical model, the main ones that change over time are the tuyere-insertion temperature, the furnace heat index, the amount of solution loss carbon, the top gas temperature of the furnace, the gas temperature at the tip of the tuyere, and the blast volume per unit. These input variables are operating variables of the blast furnace and are values measured by sensors installed in the blast furnace or obtained from known conversion formulas based on the measured values. Here, the operating variables of the blast furnace are parameters set or measured in relation to the operation of the blast furnace. The input to the statistical model may include at least one of these operating variables. The tuyere-insertion temperature, the top gas temperature of the furnace, and the gas temperature at the tip of the tuyere are the temperatures of the gas at the inserted part of the tuyere, the top gas of the furnace, and the tip of the tuyere, respectively. The furnace heat index is the amount of heat supplied into the blast furnace. The blast volume per unit is the blast volume used for producing one ton of hot metal.

[0028] Also, the output variable of the statistical model is the Si concentration. It is possible to calculate the Si concentration that changes every moment using the statistical model. The current value and predicted value of the Si concentration can be calculated (predicted) using the statistical model. Fig. 2 shows the result of predicting the hot metal Si two hours later using the statistical model (the LW-PLS model in this embodiment), and good prediction accuracy is obtained.

[0029] Here, regarding how many hours in the future the predicted value is to be obtained, it is necessary to flexibly change according to the operating action for controlling the Si concentration. For example, in the case of an operating action to reduce the Si concentration by adding a desiliconizing agent on the casting floor, since the immediacy for the change is high, it is sufficient to predict the current value. On the other hand, in the case of an operating action to control the heat level in the blast furnace to increase the Si concentration at the time of tapping before desiliconization, since the influence of thermal inertia is large and the immediacy is weak, prediction more than two hours ahead is essential.

[0030] In a statistical model, predictions are made based on measurement values from blast furnace sensors, so the accuracy decreases in the case of long-term predictions. As described above, in a statistical model, sensor information inside the furnace such as the tuyere injection temperature and the top gas temperature is used as input. When controlling the heat level inside the blast furnace, accurate predictions become possible only after the sensor information begins to reflect changes in the heat level. Therefore, there is a risk that the prediction accuracy will decrease when the influence of the operating parameter change has not appeared in the sensor information, such as immediately after changing the operating parameters (pulverized coal flow rate, coke ratio, etc.) in the blast furnace. That is, it is difficult to obtain sufficient accuracy with a statistical model for long-term predictions (for example, predictions more than 2 hours ahead). Therefore, it is preferable to use a physical model for long-term predictions.

[0031] The physical model is a physical model (unsteady model) capable of calculating the state inside the blast furnace (inside the furnace) in an unsteady state, which is composed of a group of partial differential equations considering physical phenomena such as ore reduction, heat exchange between ore and coke, and ore melting, in the same way as the method described in the reference (K. Takatani et al. ISIJ International, Vol. 39 (1999), pp. 15). The unsteady state includes, for example, the occurrence of phenomena such as blow-through and scaffolding.

[0032] As shown in Figure 3, the main input variables that change with time among the input variables given to the physical model are the blast volume flow rate, blast oxygen flow rate, pulverized coal flow rate, coke ratio, blast moisture, and blast temperature. These input variables are operating variables of the blast furnace. The input to the physical model may include at least one of these operating variables. The blast volume flow rate, blast oxygen flow rate, and pulverized coal flow rate are the flow rates of air, oxygen, and pulverized coal sent to the blast furnace, respectively. The coke ratio is the coke ratio at the top of the furnace and is the weight of coke used per ton of hot metal production. The blast moisture is the humidity of the air sent to the blast furnace. The blast temperature is the temperature of the air sent to the blast furnace.

[0033] In addition, the main output variables of the physical model are gas utilization rate, amount of solution loss carbon, ratio of reducing agent, pig iron production rate, pig iron production temperature, and Si concentration. It is possible to calculate the Si concentration that changes moment by moment using the physical model. The time interval for this calculation is not particularly limited, but in this embodiment, it is 1 hour. The time difference between "t + 1" and "t" in the formula of the physical model described later is 1 hour in this embodiment. In this embodiment, the physical model is a three-dimensional unsteady model that can estimate the three-dimensional temperature distribution in the furnace and the like. However, the form of the physical model is not limited to a three-dimensional unsteady model.

[0034] The physical model can be represented by the following formulas (1) and (2).

[0035]

Number

[0036] Here, x(t) is a state variable calculated within the physical model. The state variables are, for example, the temperature of coke, the temperature of iron, the degree of oxidation of ore, the dropping rate of raw materials, and the like. y(t) is the control variable, which is the Si concentration. u(t) is the above input variable, which is a variable that can be operated by an operator who operates the blast furnace. That is, the input variables are blast volume BV(t), blast oxygen flow rate BVO(t), pulverized coal injection flow rate PCI(t), coke ratio CR(t), blast moisture BM(t), and blast temperature BT(t). It can be expressed as u(t) = (BV(t), BVO(t), PCI(t), CR(t), BM(t), BT(t)). By repeatedly calculating formulas (1) and (2), the future Si concentration can be predicted.

[0037] In this embodiment, depending on "how many hours ahead the predicted value is to be obtained", the Si concentration is predicted by selectively using a statistical model and a physical model, and a highly accurate prediction is combined and presented to, for example, an operator or the like. FIG. 4 is an example of a screen showing the predicted value of the Si concentration predicted in this way to a milling operator and a steelmaking operator. The screen of FIG. 4 is a guidance screen that enables an operator to take appropriate operation actions according to the predicted Si concentration. In the example of FIG. 4, from the tuyere embedding temperature to the tuyere tip gas temperature indicates the operation result of the blast furnace, and from the blast volume to the blast oxygen volume indicates the operation variables that can be operated by the operator in the operation of the blast furnace. The horizontal axis indicates time, and zero corresponds to the current time. Also, in the example of FIG. 4, the target value of the Si concentration is 0.5 [wt%]. In this embodiment, a statistical model is used for the predicted value up to 2 hours ahead, and a physical model is used for the predicted value 3 hours or more ahead. That is, a statistical model is used for predicting the current value of the Si concentration and the predicted value up to a predetermined time, and a physical model is used for predicting the predicted value of the Si concentration after the predetermined time. An example of the predetermined time is 2 hours. Here, the predetermined time may be determined based on the time when the prediction accuracy of the statistical model decreases compared to the prediction accuracy of the physical model. For example, the predetermined time may be determined by comparing the prediction accuracies of the statistical model and the physical model through experiments using the operation performance data (such as past measurement data) of the operation. For example, when the prediction target time is extended to 2.5 hours later in the experiment and the prediction accuracy of the statistical model decreases compared to the prediction accuracy of the physical model, the predetermined time may be determined to be 2 hours.

[0038] Conventionally, steelmaking operators often obtained only the estimated information of the current Si concentration based on the measured values. By presenting at least the predicted trend of the Si concentration as shown in Figure 4 to the steelmaking operator, it becomes possible to contact the steelmaking operator from the steelmaking operator to the steel milling operator at an early timing with appropriate operation actions suitable for adjusting the input amount of the desiliconizing material and the Si concentration. Although the steel milling operator actually executes the operation actions, by the steelmaking operator who grasps the tapping plan etc. checking the screen of Figure 4 and contacting the steel milling side, the cooperation that was impossible with the prediction and judgment limited only to the conventional steel milling side becomes possible. Here, it may be possible for the steel milling operator and the steelmaking operator to simultaneously view the predicted trend of the same Si concentration by the statistical model and the physical model while being in different locations. That is, the predicted trend of the Si concentration as shown in Figure 4 may be simultaneously displayed on a display device visible to the steel milling operator and the steelmaking operator. At this time, strong cooperation becomes possible by sharing the same information. As a result, it becomes possible to reduce the steelmaking cost.

[0039] As a specific operation, for example, it is conceivable to change the target value of the future Si concentration. This is because the action of controlling the Si concentration, which will be described later, is determined based on the deviation between the target value of the Si concentration and the predicted value of the Si concentration. The target value of the Si concentration is determined in relation to the blast furnace operation with reference to the tapping plan in steelmaking, etc. The tapping plan is determined in multiple stages with different time units, such as, for example, based on the monthly production plan or weekly production plan, etc., making a tapping plan up to about 24 hours ahead for each day and then finalizing the ultimate tapping plan every about 6 hours. Since the blast furnace has a long time constant (for example, it may be 8 hours or more), the target value of the Si concentration is tentatively determined in consideration of the operating conditions based on the Si concentration required by the steelmaking plant obtained from the tapping plan every 24 hours. Then, based on the finally determined tapping plan, the required Si concentration is determined. In the steelmaking plant, it is possible to determine the target value of the Si concentration by obtaining the information on the changed Si concentration from the steelmaking plant (as a specific example, the corrected target value of the Si concentration and the future time or time zone to be corrected, which will be described later). Also, the tapping order in steelmaking may be changed several hours in advance due to the tapping of emergency materials or troubles, etc. In such a case as well, it can be dealt with by changing the target value of the Si concentration, etc.

[0040] Also, Si in the hot metal becomes a heat source in the steelmaking stage. Therefore, for hot metal with an Si concentration below the target value, additional processes that lead to an increase in costs, such as the charging process of Si alloy or other auxiliary raw materials that become heat sources, occur to ensure the heat source. Also, in some cases, the allocation of the target material becomes impossible, which may lead to a reexamination of the allocation or an increase in inventory. If it becomes possible to change the target value of the Si concentration, etc., such an increase in costs can be prevented, and the steelmaking cost can be reduced.

[0041] As an operation action for controlling the Si concentration, when the Si concentration is reduced by adding a desilication material, the details of the adjustment amount are described below. Based on the sensitivity between the unit consumption of the desilication material and the reduction amount of the Si concentration, when it is predicted that the Si concentration is excessive with respect to the target value, the input amount of the desilication material may be adjusted so that the target value and the predicted value of the Si concentration match. The adjustment amount is Δu in the following formula (3).

[0042]

Number

[0043] Here, α is a relaxation coefficient. y ref is the target value of the Si concentration. y pre (t) is the predicted value of the Si concentration after t hours by the statistical model or physical model calculated by the above method. S u (t) is the change amount of the Si concentration after t hours when the desilication material is operated by a unit amount, and is obtained by a step response test in an actual furnace. Here, although it is assumed that the target value of the Si concentration is constant in formula (3), when the target value is changed, y ref (t) may be given as the target value of the Si concentration after t hours. Here, since the operation action of adding the desilication material has high immediacy, t = 0 may be set.

[0044] For example, when the Si concentration still fails to reach the target even if the addition of the desilication material is stopped, an operation action different from the desilication material addition action is required. As an operation action for controlling the Si concentration, when controlling the heat level in the blast furnace to increase the Si concentration at the time of tapping before desilication, the details of the adjustment amount are described below. For example, when the blast moisture is used as the operation variable to be adjusted, the adjustment amount of the blast moisture is ΔBM in the following formula (4).

[0045]

Number

[0046] Here, β is a relaxation coefficient. SBM (t) is the change amount of the Si concentration after t hours when the blowing moisture is operated by a unit amount, and is obtained by a step response test in an actual furnace or a step response calculation based on a physical model. Not only the blowing moisture, but also the same control can be performed using, for example, the blowing temperature or the pulverized coal flow rate. Here, although the target value of the Si concentration is assumed to be constant in Equation (4), when the target value is changed, y ref (t) may be given as the target value of the Si concentration after t hours. Here, when these operating variables are used, it is desirable to set t = 2 to 8. Also, since the response time varies depending on the operating variable, the value of t is determined according to the selected operating variable.

[0047] The calculation of the adjustment amount according to Equation (3) or Equation (4) may be executed according to an instruction from an operator, but in this embodiment, it is automatically executed when the predicted value or the current value of the Si concentration greatly deviates from the target value. Here, "greatly deviates from the target value" may be determined by the magnitude of the value obtained by subtracting the target value from the predicted value and the current value exceeding a threshold value. The threshold value may be determined based on, for example, the magnitude of the deviation from the target value allowed in the production of hot metal.

[0048] FIG. 5 is a diagram showing a configuration example of the Si concentration prediction device 10 according to an embodiment. As shown in FIG. 5, the Si concentration prediction device 10 includes a storage unit 11, a prediction unit 12, and an adjustment amount calculation unit 13. The Si concentration prediction device 10 acquires operation variables including measured values (sensor information), and predicts the Si concentration using the above statistical model and physical model. Further, the Si concentration prediction device 10 may acquire various instructions including, for example, an instruction to start prediction processing. The Si concentration prediction device 10 may present the predicted predicted value and the current value to the operator, and in this case, it also functions as an operation guidance device. When functioning as an operation guidance device, the Si concentration prediction device 10 causes the display unit 30 to display the predicted future predicted value of the predicted Si concentration, the current value, and the like. The display unit 30 may be a display device such as a liquid crystal display (Liquid Crystal Display) or an organic EL panel (Organic Electro-Luminescence Panel). Here, the display unit 30 may be a display device included in the terminal device 40 used by the operator. The operator includes operators of milling and steelmaking. A screen such as FIG. 4 may be displayed on the display unit 30. The terminal device 40 may be a portable terminal such as a smartphone or a tablet. An operation guidance system may be configured including the operation guidance device (Si concentration prediction device 10) and the terminal device 40. In the operation guidance system, the terminal device 40 may include an acquisition unit 41, an input unit 42, a transmission unit 43, and a display unit 30. The terminal-to-terminal communication unit 44 will be described later. The acquisition unit 41 may acquire the predicted value, the current value, and the adjustment amount of the operation variable in the operation action output from the operation guidance device. The input unit 42 is realized by an input device such as a keyboard or a pointing device, and a correction value of the target value of the Si concentration and the future time or time zone to be corrected are input. For example, input is performed when an operator of a steelmaking plant who has confirmed the change in the predicted value of the future Si concentration and the change status of the tapping plan requests the operator of the milling plant to change the target value of the future Si concentration. The request to change the target value of the Si concentration may be made by another communication means without passing through the system. Hereinafter, the changed target value may be referred to as a corrected target value depending on the request for change.The transmission unit 43 transmits the input corrected target value of the Si concentration and the future time or time zone to be corrected to the operation guidance device (Si concentration prediction device 10). Further, as described above, the display unit 30 may display the predicted value, the current value, and the adjustment amount acquired by the acquisition unit 41. The display unit 30 may further display the corrected target value and the future time or time zone to be corrected. Also, it may be possible to transmit and receive messages between a plurality of terminal devices 40. In this case, the display unit 30 has a function of displaying the transmitted and received messages. As shown in FIG. 5, the plurality of terminal devices 40 are provided with an inter-terminal communication unit 44, and it may be possible to transmit and receive information such as a request to change the target value of the Si concentration, the change time zone, and the corrected target value between the plurality of terminal devices 40. For example, information regarding the correction of the target value of the Si concentration may be transmitted and received between the terminal device 40 used by the operator of steelmaking and the terminal device 40 used by the operator of steel milling by their respective inter-terminal communication units 44, and information sharing may be performed in real time. Also, using such a function, the corrected target value of the Si concentration may be set based on the input information (for example, the above request to change the target value of the Si concentration) from the operator of steelmaking (operator other than steel milling). Here, the acquisition unit 41 may be realized by, for example, a processor (arithmetic processing unit) of the terminal device 40. Also, the number of terminal devices 40 is not limited to one, and there may be a plurality as in the above example. For example, the terminal device 40 may be installed at the respective resident locations of the operators of steelmaking and steel milling and may be communicably connected to the operation guidance device by a network such as a LAN (Local Area Network) or other communication means. Also, the inter-terminal communication unit 44 may be a communication means using a network, but is not limited to a specific one as long as it can transmit and receive information regarding the corrected target value of the Si concentration to and from another terminal device.

[0049] The components of the Si concentration prediction device 10 will be described below. The storage unit 11 stores a statistical model and a physical model that take the operating variables of the blast furnace as input and output the Si concentration. Further, the storage unit 11 stores a program and data related to the prediction of the Si concentration. The storage unit 11 may include any storage device such as a semiconductor storage device, an optical storage device, and a magnetic storage device. The semiconductor storage device may include, for example, a semiconductor memory. The storage unit 11 may include a plurality of types of storage devices.

[0050] The prediction unit 12 predicts future predicted values and current values of the Si concentration using the statistical model and the physical model. When the Si concentration prediction device 10 also functions as an operation guidance device, the prediction unit 12 presents the predicted predicted values and current values to the operator by causing the display unit 30 to display them.

[0051] When the predicted value or the current value deviates significantly from the target value, the adjustment amount calculation unit 13 calculates the adjustment amount of the operating variable in the operation action. The operation action may be to reduce the Si concentration by adding a desiliconizing material, or to control the heat level in the blast furnace to increase the Si concentration at the time of tapping before desiliconization. That is, the adjustment amount calculated by the adjustment amount calculation unit 13 includes the adjustment amounts for both the case of reducing the Si concentration and the case of increasing it, and includes, for example, the adjustment amount of the input amount of the desiliconizing material. In the case of the operation action of adding a desiliconizing material, the adjustment amount calculation unit 13 may calculate the adjustment amount for the input amount of the desiliconizing material according to the above formula (3). Here, if a corrected target value is input, the adjustment amount calculation unit 13 calculates the adjustment amount using the corrected target value as the target value. In the case of the operation action of controlling the heat level in the blast furnace, the adjustment amount calculation unit 13 may calculate the adjustment amount for the blast moisture, blast temperature, or pulverized coal flow rate as in the above formula (4). For example, a computer that manages the production of hot metal may automatically change the production conditions of hot metal based on the calculated adjustment amount. Here, when the Si concentration prediction device 10 also functions as an operation guidance device, the adjustment amount calculation unit 13 may present the calculated adjustment amount to the operator by displaying it on the display unit 30. Such a presentation to the operator (operation guidance) can be executed as a part of the production method for producing hot metal or steel.

[0052] The Si concentration prediction device 10 may be realized by a computer such as a process computer that controls the operation of a blast furnace or the production of hot metal. The computer includes, for example, a memory and a hard disk drive (storage device), a CPU (processing device), and the like. Various functions are realized by the organic cooperation of hardware such as the CPU and memory, the OS, and necessary application programs. The storage unit 11 may be realized by, for example, a storage device. The prediction unit 12 and the adjustment amount calculation unit 13 may be realized by, for example, a processing device.

[0053] FIG. 6 is a flowchart showing a Si concentration prediction method according to an embodiment. The Si concentration prediction apparatus 10 predicts the Si concentration according to the flowchart shown in FIG. 6. The Si concentration prediction method shown in FIG. 6 may be executed as part of a method for manufacturing hot metal.

[0054] The prediction unit 12 predicts future predicted values and current values of the Si concentration using a statistical model and a physical model (step S1, prediction step). The prediction unit 12 may cause the predicted predicted values and current values to be presented to the operator.

[0055] When the predicted value or the current value of the Si concentration greatly deviates from the target value (Yes in step S2), the adjustment amount calculation unit 13 calculates the adjustment amount of the operation variable in the operation action as described above (step S3, adjustment amount calculation step). In the adjustment amount calculation step, if a corrected target value is input, the adjustment amount calculation unit 13 calculates the adjustment amount of the operation variable using the corrected target value as the target value. The process of the Si concentration prediction method may further include a step (input step) in which a corrected target value is input via the input unit 42 and the transmission unit 43. A series of processes of the Si concentration prediction method may be repeatedly executed at a predetermined cycle, or may be executed again, for example, when the target value is changed. The adjustment amount calculation unit 13 may cause the calculated adjustment amount to be presented to the operator. Here, when neither the predicted value nor the current value of the Si concentration deviates from the target value (No in step S2), it is not necessary to calculate the adjustment amount.

[0056] As described above, the Si concentration prediction method, the operation guidance method, the Si concentration prediction apparatus 10, the operation guidance apparatus, the operation guidance system, and the terminal device 40 according to the present embodiment can accurately predict the Si concentration by properly using the statistical model and the physical model according to the time of the prediction target. Further, when there is a request to change the Si concentration from the steelmaking plant, by changing the target value of the Si concentration, it is possible to appropriately calculate the operation action (adjustment amount of the operation variable) according to the predicted values of the statistical model and the physical model in the steelmaking plant. Therefore, even in the operation involving the steelmaking plant and the steelmaking plant, it is possible to prevent the deviation from the target value of the Si concentration.

[0057] Although the embodiments according to the present disclosure have been described based on the drawings and examples, it should be noted that those skilled in the art can easily make various modifications or corrections based on the present disclosure. Therefore, it should be noted that these modifications or corrections are included in the scope of the present disclosure. For example, functions included in each component or each step can be rearranged so as not to be logically contradictory, and a plurality of components or steps can be combined into one or divided. The embodiments according to the present disclosure can also be realized as a program executed by a processor included in a device or a storage medium storing the program. It should be understood that these are also included in the scope of the present disclosure.

[0058] The configuration of the Si concentration prediction device 10 shown in FIG. 5 is an example. The Si concentration prediction device 10 does not necessarily include all of the components shown in FIG. 5. Further, the Si concentration prediction device 10 may include components other than those shown in FIG. 5. Further, the operation guidance system includes an operation guidance device (Si concentration prediction device 10) and a terminal device 40, but the components included in each are not limited to the examples shown in FIG. 5. For example, a configuration in which at least one of the prediction unit 12 and the adjustment amount calculation unit 13 is included in the terminal device 40 is possible. Therefore, the operation guidance system may be configured such that the operation guidance device or the terminal device 40 includes a storage unit 11, a prediction unit 12, an adjustment amount calculation unit 13, an acquisition unit 41, an input unit 42, a transmission unit 43, and a display unit 30.

Explanation of Reference Numerals

[0059] 10 Si concentration prediction device 11 Storage unit 12 Prediction unit 13 Adjustment amount calculation unit 30 Display unit 40 Terminal device 41 Acquisition unit 42 Input unit 43 Transmission unit 44 Inter-terminal communication unit

Claims

1. A prediction step of predicting a future predicted value and a current value of the Si concentration using a statistical model and a physical model having an input of an operational variable of a blast furnace and an output of the Si concentration; and an adjustment amount calculation step of calculating, when the predicted value or the current value significantly deviates from a target value, an adjustment amount of an amount of desiliconization material fed in an operational action of feeding a desiliconization material to reduce a Si concentration, or an adjustment amount of the operational variable in an operational action of controlling a heat level in a blast furnace to increase a Si concentration at the time of tapping before desiliconization, In the prediction step, the statistical model is used to predict the current value and the predicted value up to a predetermined time, and the physical model is used to predict the predicted value beyond the predetermined time, the predetermined time is determined based on a time period during which the prediction accuracy of the statistical model becomes lower than the prediction accuracy of the physical model; A Si concentration prediction method in which u(t) is (BV(t), BVO(t), PCI(t), CR(t), BM(t), BT(t)), using time as t, x(t) as a state variable calculated in the physical model, y(t) as a Si concentration, blast flow rate BV(t), blast oxygen flow rate BVO(t), pulverized coal flow rate PCI(t), coke ratio CR(t), blast moisture BM(t), and blast temperature BT(t), and the physical model is expressed by the following equations (1) and (2). [0010]

2. The Si concentration prediction method according to claim 1, wherein the operational variables that are inputs to the statistical model include at least one of tuyere embedding temperature, furnace heat index, solution loss carbon amount, furnace top gas temperature, tuyere tip gas temperature, and blast volume consumption rate.

3. 3. An operation guidance method comprising the step of having an operator present the predicted value and the current value predicted by the Si concentration prediction method according to claim 1 or 2.

4. An input step of inputting a corrected target value of the Si concentration, 4. The operation guidance method according to claim 3, wherein the adjustment amount calculation step calculates an adjustment amount for the operation variable using the corrected target value input in the input step as the target value.

5. The operation guidance method according to claim 4 , wherein the corrected target value is set based on input information from an operator other than an ironmaking operator.

6. A storage unit that stores a statistical model and a physical model in which the operation variables of the blast furnace are input and the Si concentration is output; a prediction unit that predicts a future predicted value and a present value of the Si concentration using the statistical model and the physical model; an adjustment amount calculation unit that calculates, when the predicted value or the current value significantly deviates from a target value, an adjustment amount of an amount of desiliconization material fed in an operational action of feeding a desiliconization material to reduce a Si concentration, or an adjustment amount of the operational variable in an operational action of controlling a heat level in a blast furnace to increase a Si concentration at the time of tapping before desiliconization, the prediction unit uses the statistical model to predict the current value and the predicted value up to a predetermined time, and uses the physical model to predict the predicted value beyond the predetermined time, the predetermined time is determined based on a time period during which the prediction accuracy of the statistical model becomes lower than the prediction accuracy of the physical model; A Si concentration prediction device in which u(t) is (BV(t), BVO(t), PCI(t), CR(t), BM(t), BT(t)), using time as t, x(t) as a state variable calculated in the physical model, y(t) as Si concentration, blast flow rate BV(t), blast oxygen flow rate BVO(t), pulverized coal flow rate PCI(t), coke ratio CR(t), blast moisture BM(t), and blast temperature BT(t), and the physical model is expressed by the following equations (1) and (2). [0025]

7. A storage unit that stores a statistical model and a physical model in which the operation variables of the blast furnace are input and the Si concentration is output; a prediction unit that predicts a future predicted value and a current value of the Si concentration by using the statistical model and the physical model, and displays the predicted value and the current value to an operator; an adjustment amount calculation unit that calculates, when the predicted value or the current value significantly deviates from a target value, an adjustment amount of an amount of desiliconization material fed in an operational action of feeding a desiliconization material to reduce a Si concentration, or an adjustment amount of the operational variable in an operational action of controlling a heat level in a blast furnace to increase a Si concentration at the time of tapping before desiliconization, the prediction unit uses the statistical model to predict the current value and the predicted value up to a predetermined time, and uses the physical model to predict the predicted value beyond the predetermined time, the predetermined time is determined based on a time period during which the prediction accuracy of the statistical model becomes lower than the prediction accuracy of the physical model; An operation guidance device, in which time is represented by t, x(t) is a state variable calculated in the physical model, y(t) is Si concentration, blast flow rate BV(t), blast oxygen flow rate BVO(t), pulverized coal flow rate PCI(t), coke ratio CR(t), blast moisture BM(t), and blast temperature BT(t), and u(t) is (BV(t), BVO(t), PCI(t), CR(t), BM(t), BT(t)), and the physical model is expressed by the following equations (1) and (2). [0030] 【number】

8. An operation guidance device and a terminal device, The operation guidance device or the terminal device is A storage unit that stores a statistical model and a physical model in which the operation variables of the blast furnace are input and the Si concentration is output; a prediction unit that predicts a future predicted value and a present value of the Si concentration using the statistical model and the physical model, and outputs the predicted value and the present value; an adjustment amount calculation unit that calculates an adjustment amount of an amount of desiliconization material fed in an operational action of feeding a desiliconization material to reduce a Si concentration, or an adjustment amount of the operational variable in an operational action of controlling a heat level in a blast furnace to increase a Si concentration at the time of tapping before desiliconization, when the predicted value or the current value significantly deviates from a target value, and outputs the adjustment amount; an acquisition unit that acquires the predicted value, the current value, and the adjustment amount; an input section for inputting a correction target value of the Si concentration and a future time or time period to be corrected; a transmission unit that transmits the input correction target value and the future time or time period to be corrected; a display unit that displays the acquired predicted value, the current value, the adjustment amount, and the correction target value, and a future time or time period to be corrected; the prediction unit uses the statistical model to predict the current value and the predicted value up to a predetermined time, and uses the physical model to predict the predicted value beyond the predetermined time, the predetermined time is determined based on a time period during which the prediction accuracy of the statistical model becomes lower than the prediction accuracy of the physical model; An operation guidance system in which time is represented by t, x(t) is a state variable calculated in the physical model, y(t) is Si concentration, blast flow rate BV(t), blast oxygen flow rate BVO(t), pulverized coal flow rate PCI(t), coke ratio CR(t), blast moisture BM(t), and blast temperature BT(t), and u(t) is (BV(t), BVO(t), PCI(t), CR(t), BM(t), BT(t)), and the physical model is expressed by the following equations (1) and (2). [0045]

9. an acquisition unit that acquires, when a future predicted value and a current value of the Si concentration predicted by the Si concentration prediction device according to claim 6, which uses a statistical model and a physical model in which an operational variable of a blast furnace is input and an Si concentration is output, an adjustment amount of an amount of the desiliconization material input in an operational action of lowering the Si concentration by inputting a desiliconization material, or an adjustment amount of the operational variable in an operational action of controlling a heat level in a blast furnace to increase the Si concentration at the time of tapping before desiliconization is calculated in a case where the predicted value or the current value significantly deviates from a target value, the predicted value, the current value, and the adjustment amount; an input section for inputting a correction target value of the Si concentration and a future time or time period to be corrected; a transmission unit that transmits the input correction target value and the future time or time period to be corrected; a display unit that displays the acquired predicted value, the current value, the adjustment amount, and the correction target value, as well as a future time or time zone to be corrected.

10. The terminal device according to claim 9 , further comprising an inter-terminal communication unit for transmitting and receiving information regarding the correction target value to and from another terminal device.

Images