Blast furnace operation methods

Abstract

The desired ore layer thickness ratio is calculated with high accuracy, using multiple operational indicators as predetermined targets. [Solution] The method for operating a blast furnace includes the steps of: obtaining an operation command to bring an operational indicator showing the state of the blast furnace within a target range; and calculating target operating conditions, which are operating conditions that bring the difference between the operational indicator after changing the operating conditions and the target operational indicator included in the operation command within a predetermined range, based on an operational indicator estimation model that takes operating conditions as input and operational indicators as output. The operation command includes target operating indicators for two or more operational indicators, where the target operating indicator for one of the two or more operational indicators maximizes, minimizes, or sets the operational indicator to a predetermined value, and the target operating indicator for the other of the two or more operational indicators sets the target operating indicator to a predetermined value, greater than or equal to a predetermined value, or less than or equal to a predetermined value; and the operating conditions include the distribution of ore layer thickness ratios.

Description

Technical Field

[0001] This disclosure relates to an operating method for a blast furnace.

Background Art

[0002] Conventionally, when operating a blast furnace, operating conditions for bringing the state of the blast furnace to an ideal state may be defined. As one of the operating conditions, for example, there is an ore layer thickness ratio.

[0003] In the blast furnace, ore layers and coke layers are alternately stacked, and the distribution shape of these ore layers and coke layers has a very large influence on the operation of the blast furnace. The distribution shape of the ore layer and coke layer charged into the blast furnace is controlled by the layer thickness ratio between the ore layer and the coke layer in the radial direction of the blast furnace. The ore layer thickness ratio means the ratio of the thickness of the ore layer to the total thickness of the thickness of the ore layer and the coke layer. The ore layer thickness ratio changes along the radial direction of the blast furnace, and even within the furnace cross-section (for example, a circular cross-section viewed from the furnace top), the layer thickness ratio varies depending on the position, so it is treated as a distribution within the cross-section.

[0004] For example, Patent Document 1 discloses a technique for defining the range of a desirable ore layer thickness ratio in the furnace center region, the furnace middle region, and the furnace peripheral region, and setting the ore layer thickness ratio within the defined desirable range. Patent Document 1 measures the ore layer thickness ratio using a profile meter, and if necessary, changes the charging method from the charging chute to set the ore layer thickness ratio within the defined desirable range.

[0005] For example, Patent Document 2 discloses a technique for measuring whether the ore layer thickness ratio is within a certain range using a characteristic value selected from the group consisting of a maximum value, a minimum value, and an average value for each target location, and changing the charging method to set the ore layer thickness ratio within a certain range when the ore layer thickness ratio does not fall within the certain range.

Prior Art Documents

Patent Documents

[0006] [Patent Document 1] Japanese Patent Publication No. 2017-95761 [Patent Document 2] Japanese Patent Publication No. 2021-113341 [Overview of the project] [Problems that the invention aims to solve]

[0007] Patent documents 1 and 2 define the ideal ore layer thickness ratio distribution based on empirical rules that aggregate knowledge of changes in blast furnace conditions due to changes in operating conditions. However, the ore layer thickness ratio distribution defined based on empirical rules may be limited to local optimal solutions.

[0008] If operating conditions such as the ore layer thickness ratio are defined based on empirical rules, and the operating conditions are not significantly changed in the same blast furnace, then the operating conditions defined based on empirical rules may be the ideal operating conditions.

[0009] However, operating conditions defined in this way based on empirical rules are not necessarily ideal when operating conditions are significantly changed or when operating in other blast furnaces. When operating conditions are significantly changed or when operating in other blast furnaces, different operating conditions may be ideal, and there was a problem in that optimality could not be guaranteed.

[0010] The purpose of this disclosure is to provide a blast furnace operation method that can accurately calculate a desired ore layer thickness ratio that sets multiple operational indicators to predetermined targets. [Means for solving the problem]

[0011] [1] A step of obtaining an operational command to bring the operational indicators showing the state of the blast furnace while the blast furnace is in operation within a target range, A step of calculating target operating conditions, which are the operating conditions after changing the operating conditions, and which are the operating conditions included in the operation command, based on an operation indicator estimation model that takes operating conditions, which are the conditions for operating the blast furnace, as input and the operation indicator as output, so that the difference between the operation indicator after changing the operating conditions and the target operation indicator included in the operation command is within a predetermined range. Includes, The aforementioned operational command includes the aforementioned target operational indicators for two or more of the aforementioned operational indicators, The target operating indicator for one of the two or more operating indicators is to maximize, minimize, or set the operating indicator to a predetermined value. The target operating indicator for the other operating indicator among the two or more operating indicators is such that the target operating indicator is set to a predetermined value, greater than or equal to a predetermined value, or less than or equal to a predetermined value. The aforementioned operating conditions include the distribution of ore layer thickness ratios, and the method of operating a blast furnace.

[0012] [2] The method for operating a blast furnace according to [1], further comprising the step of presenting the calculated target operating conditions to the operator, or the step of automatically changing the operating conditions based on the calculated target operating conditions.

[0013] [3] The method of operating a blast furnace according to [1] or [2] above, wherein the operating indicators include at least one of the aeration resistance index, gas utilization rate, heat load, and solution loss carbon amount.

[0014] [4] A method for operating a blast furnace according to any one of [1] to [3] above, wherein the operating conditions further include at least one of the following: oxygen enrichment rate, blast flow rate, iron production rate, pulverized coal ratio, heat flow ratio, sintered ore ratio, coke lump ratio, coke strength, sintered ore strength, sintered ore moisture content, blast temperature, and blast furnace gas flow rate distribution.

[0015] [5] The distribution of the ore layer thickness ratio, the distribution of the gas flow rate in the blast furnace, and the distribution of the heat flow ratio have two or more values in the radial direction or within the cross-section inside the blast furnace, or are characteristic quantities indicating the distribution in the radial direction or the cross-section inside the blast furnace calculated from the two or more values. The operation method of the blast furnace according to any one of [1] to [4] above.

[0016] [6] The characteristic quantity is a principal component score obtained by dimension compression of the distribution in the radial direction or within the cross-section inside the blast furnace by principal component analysis (PCA). The operation method of the blast furnace according to any one of [1] to [5] above.

[0017] [7] The characteristic quantity is characterized by using the first to Nth principal components (N is an integer including 5) with a cumulative contribution rate of 70% or more and 90% or less. The operation method of the blast furnace according to any one of [1] to [6] above.

[0018] [8] The operation index estimation model is a model constructed based on the past operation performance of the blast furnace. The operation method of the blast furnace according to any one of [1] to [7] above.

Advantages of the Invention

[0019] According to the operation method of the blast furnace according to the present disclosure, a desired ore layer thickness ratio that sets a plurality of operation indexes to a predetermined target can be accurately calculated.

Brief Description of the Drawings

[0020] [Figure 1] It is a diagram showing an example of an operation condition control device for executing the operation method of the blast furnace according to an embodiment of the present disclosure. [Figure 2] It is a flowchart showing an example of the operation method of the blast furnace according to an embodiment of the present disclosure. [Figure 3] It is a diagram showing an example of the estimation result of the ventilation resistance index. [Figure 4] It is a diagram showing an example of the ore layer thickness ratio calculated based on the operation index estimation model. [Figure 5]This figure shows an example of the air permeability resistance index when the ore layer thickness ratio is changed. [Figure 6] This figure shows an example of gas utilization rates when the ore layer thickness ratio is changed. [Figure 7] This figure shows an example of the distribution shape of the first to fifth principal components of the gas flow rate distribution inside the blast furnace. [Figure 8] This figure shows an example of the prediction result of the airflow resistance index using the heat flux ratio distribution. [Figure 9] This figure shows an example of the prediction results for the airflow resistance index using the gas flow rate distribution. [Figure 10] This figure shows an example of the revised ore layer thickness ratio distribution calculated based on the operational instructions. [Figure 11A] This figure shows an example of the air permeability resistance index when the ore layer thickness ratio is changed. [Figure 11B] This figure shows an example of gas utilization rates when the ore layer thickness ratio is changed. [Modes for carrying out the invention]

[0021] The embodiments of this disclosure will be described below with reference to the drawings.

[0022] Figure 1 shows an example of an operating condition control device 10 that performs a blast furnace operating method according to one embodiment of the present disclosure.

[0023] The operating condition control device 10 is a device that controls the operating conditions, which are the conditions under which the blast furnace 1 is operated. By controlling the operating conditions, the operating condition control device 10 can bring the operating indicators, which show the state of the blast furnace 1, within a target range.

[0024] The operating conditions for operating blast furnace 1 include the distribution of the ore layer thickness ratio. Blast furnace 1 consists of alternating layers of ore and coke, and the ore layer thickness ratio is defined as the ratio of the thickness of the ore layer to the sum of the thicknesses of the ore and coke layers, as shown in equation (1) below.

number

[0025] The operating conditions, in addition to the ore layer thickness ratio, further include at least one of the following: oxygen enrichment rate, blast flow rate, iron production rate, pulverized coal ratio, heat flow ratio, sintered ore ratio, small coke lump ratio, coke strength, sintered ore strength, sintered ore moisture content, blast temperature, and blast furnace gas flow rate distribution.

[0026] Here, the heat flow ratio is the ratio of the heat capacities of the solid and gas in the furnace. The heat flow ratio includes heat flow ratio distributions obtained by measuring the ore descent rate and gas flow rate in the furnace, heat flow ratio distributions obtained by reaction calculations in the blast furnace, and heat flow ratio distributions estimated from the gas temperature distribution. Furthermore, the gas flow rate distribution in the blast furnace includes gas flow rate distributions obtained using flow meters installed in the blast furnace, gas flow rate distributions obtained by blast furnace simulations, and gas flow rate distributions estimated using thermal balance.

[0027] Furthermore, the distribution of ore layer thickness ratio, the blast furnace gas flow rate distribution, and the heat flux ratio distribution may have two or more values ​​(individual values) within the blast furnace 1 in the radial direction or cross-section. Alternatively, the ore layer thickness ratio and the blast furnace gas flow rate distribution may be feature quantities calculated from two or more values ​​in the radial direction or cross-section. The feature quantities may be, for example, the mean, gradient, or standard deviation. Alternatively, the feature quantity may be a single index such as the terrace length, which is the length of the portion that continues almost horizontally within a certain height range in the charge distribution profile (radial height distribution viewed from the top of the furnace).

[0028] Alternatively, the features may be principal component scores (values ​​projected onto the principal component axes) obtained by dimensionality reduction of the radial or cross-sectional distribution inside the blast furnace using principal component analysis (PCA). Specifically, PCA is applied to high-dimensional data in which each distribution is represented by values ​​at multiple measurement points, and multiple principal components that can represent the in-furnace distribution shape are extracted. Then, by using the principal component scores for each sample as features, the main variation patterns of the distribution shape can be represented with a small number of explanatory variables.

[0029] This allows the model to be represented by a small number of principal component scores, even when there are, for example, 20 measurement points in the radial direction, rather than using these directly as explanatory variables. This prevents the model from becoming overly complex and presenting unnatural distributions. Conventional methods may require extremely different values ​​for adjacent points, potentially leading to unrealistic results or distributions that differ significantly from past operational performance. However, this method uses principal components that reflect the trend of the distribution shape based on training data, allowing past distribution shapes to be reproduced with fewer explanatory variables.

[0030] Furthermore, since standardization (mean 0, variance 1) is performed during PCA derivation, the principal component score represents the multiplier of the standard deviation in the training data. Therefore, by constraining the principal component score to within ±2.0, it is possible to guarantee that the distribution falls within the 95% range of the training data, thereby maintaining realistic operating conditions.

[0031] Furthermore, it is desirable to select principal components from the first to the Nth principal components whose cumulative contribution rate (the cumulative value of the proportion of variance explained by each principal component) is within a predetermined range (e.g., 70% to 90%). If the cumulative contribution rate is less than 70%, it may not be possible to explain most of the variance of the original data, and the main features of the distribution shape may be lost. Therefore, in order to adequately reflect the distribution pattern that affects the operational indicators, it is necessary to retain at least about 70% of the variance. On the other hand, if the cumulative contribution rate is set excessively high (e.g., 95% or more), the number of principal components will increase too much, leading to an excessive number of explanatory variables, which can cause overfitting or an increased computational load. The portion exceeding 90% often represents noise or local fluctuations, and may not only fail to contribute to improving model accuracy but may even have a negative impact. Therefore, it is desirable to select principal components from the first to the Nth principal components whose cumulative contribution rate is between 70% and 90%.

[0032] Based on the above, when feature-quantifying in-furnace distributions (heat flux ratio distribution, ore layer thickness ratio distribution, gas flow rate distribution), using principal component analysis (PCA) allows for multidimensional compression of the overall distribution shape, representing the essential shape of the distribution. Furthermore, differences in furnace diameter and absolute value scale can be absorbed through standardization, making it easier to apply to different blast furnaces (generalization performance). In addition, since the principal component scores are dimensionally compressed using orthogonal components, the risk of multicollinearity and overfitting can be reduced when used as input (explanatory variables) for machine learning models.

[0033] The operational indicators representing the state of blast furnace 1 include at least one of the following: aeration resistance index, gas utilization rate, heat load, and solution loss carbon amount.

[0034] The operating conditions control device 10 may be a general-purpose computer such as a workstation or personal computer, or it may be a dedicated computer configured to function as the operating conditions control device 10.

[0035] The operating conditions control device 10 comprises a control unit 11, an input unit 12, an output unit 13, a storage unit 14, and a communication unit 15.

[0036] The control unit 11 includes at least one processor, at least one dedicated circuit, or a combination thereof. The processor is a general-purpose processor such as a CPU (Central Processing Unit) or GPU (Graphics Processing Unit), or a dedicated processor specialized for a specific process. The dedicated circuit is, for example, an FPGA (Field-Programmable Gate Array) or an ASIC (Application Specific Integrated Circuit).

[0037] The control unit 11 reads programs, data, etc., stored in the memory unit 14 and executes various functions.

[0038] The input unit 12 includes one or more input interfaces that detect user input and acquire input information based on user operations. The input unit 12 includes, for example, physical keys, capacitive keys, a touchscreen integrated with the display of the output unit 13, or a microphone that accepts voice input.

[0039] The output unit 13 includes one or more output interfaces for outputting information and notifying the user. The output unit 13 includes, for example, a display for outputting information as an image, a speaker for outputting information as sound, etc. The display included in the output unit 13 may be, for example, an LCD (Liquid Crystal Display), a CRT (Cathode Ray Tube) display, etc.

[0040] The storage unit 14 is, for example, a flash memory, a hard disk, or an optical memory. Part of the storage unit 14 may be located outside the operating condition control device 10. In this case, part of the storage unit 14 may be a hard disk, memory card, or the like, connected to the operating condition control device 10 via any interface.

[0041] The memory unit 14 stores programs for the control unit 11 to execute various functions, data used by those programs, and so on.

[0042] The communication unit 15 includes at least one of a communication module that supports wired communication and a communication module that supports wireless communication. The operating condition control device 10 can communicate with other devices via the communication unit 15.

[0043] Next, we will explain the blast furnace operation method executed by the operating condition control device 10.

[0044] The control unit 11 of the operating condition control device 10 acquires an operating command to bring the operating indicators, which show the state of the blast furnace 1, within a target range. The control unit 11 may acquire the operating command, for example, by receiving the input operation of the operating command by the operator via the input unit 12. Alternatively, the control unit 11 may acquire the operating command transmitted from another device via the communication unit 15, for example.

[0045] An operational instruction may include target operational indicators for two or more operational indicators. Here, a target operational indicator is the target value of the operational indicator. For example, an operational instruction may include target operational indicators for each of three operational indicators: ventilation resistance index, gas utilization rate, and heat load.

[0046] If an operational instruction includes target operational indicators for two or more operational indicators, the target operational indicator for one operational indicator may be to maximize, minimize, or bring the operational indicator to a predetermined value. Here, the predetermined value is not limited to a specific value, but may be a predetermined range centered on that specific value. Furthermore, the target operational indicators for other operational indicators may be to bring the target operational indicator to a predetermined value. Here, the predetermined value is not limited to a specific value, but may be a predetermined range centered on that specific value. Furthermore, the target operational indicators for other operational indicators may be to bring the target operational indicator to a predetermined value or to a predetermined value or to a predetermined value or less.

[0047] For example, if an operational instruction includes target operational indicators for three operational indicators—airflow resistance index, gas utilization rate, and heat load—the target operational indicator for the airflow resistance index may be to minimize the airflow resistance index. Similarly, the target operational indicator for the gas utilization rate may be to set the gas utilization rate to a predetermined value. Furthermore, the target operational indicator for the heat load may be to set the heat load to a predetermined value.

[0048] When an operational command includes the three target operational indicators described above, the command is to minimize the ventilation resistance index while maintaining the gas utilization rate and heat load at predetermined values. In this case, if the predetermined values ​​are equivalent to the current gas utilization rate and heat load, the command is to minimize the ventilation resistance index while maintaining the gas utilization rate and heat load near their current values. By controlling the operating conditions in accordance with such an operational command, the operational condition control device 10 can minimize the ventilation resistance index while maintaining the gas utilization rate and heat load at desired values.

[0049] The memory unit 14 stores an operational indicator estimation model. The operational indicator estimation model is a model that takes operating conditions as input and outputs an operational indicator. The memory unit 14 may store an operational indicator estimation model that can output multiple operational indicators, or it may store multiple operational indicator estimation models that can output one operational indicator.

[0050] If the operational indicator estimation model is capable of outputting multiple operational indicators, it may be capable of outputting, for example, three operational indicators such as the ventilation resistance index, gas utilization rate, and heat load. Alternatively, if the operational indicator estimation model is capable of outputting only one operational indicator, the storage unit 14 may store, for example, three operational indicator estimation models: one that outputs the ventilation resistance index, one that outputs the gas utilization rate, and one that outputs the heat load.

[0051] The operational indicator estimation model is a model constructed based on the past operational performance of blast furnace 1. The memory unit 14 stores the past operational performance of blast furnace 1 as a database. The memory unit 14 stores past operating conditions and past operational indicators as past operational performance. Note that the past operational performance of blast furnace 1 may be stored in a device other than the operational condition control device 10 instead of the memory unit 14.

[0052] The operational indicator estimation model may be, for example, a model constructed by machine learning based on the past operational performance of blast furnace 1. The operational indicator estimation model may be a model constructed by machine learning based on past operational performance over a predetermined period, or it may be a model that is updated each time additional operational performance data is added, by performing machine learning on that data.

[0053] When constructing an operational indicator estimation model using machine learning, methods such as linear regression, deep neural networks, decision trees, and support vector machines may be used.

[0054] When the control unit 11 receives an operation command, it calculates target operating conditions based on an operation indicator estimation model to bring the difference between the operation indicator after changing the operating conditions and the target operation indicator included in the operation command within a predetermined range. Here, the target operating conditions are the operating conditions under which changing from the current operating conditions to the target operating conditions will bring the difference between the operation indicator after changing the operating conditions and the target operation indicator included in the operation command within a predetermined range.

[0055] The control unit 11 may, for example, input the history of operating conditions included in past operating results into an operating indicator estimation model to calculate operating indicators, and set operating conditions such that the difference between the calculated operating indicators and the target operating indicators included in the operating command falls within a predetermined range as the target operating conditions.

[0056] When the control unit 11 calculates the target operating conditions, it may output the calculated target operating conditions to the output unit 13 and present them to the operator. This allows the operator, having grasped the target operating conditions, to change the current operating conditions to the target operating conditions and control the operating indicators of the blast furnace 1 to meet the operation command.

[0057] Alternatively, once the control unit 11 calculates the target operating conditions, it may automatically change the operating conditions from the current operating conditions to the target operating conditions based on the calculated target operating conditions. This allows the control unit 11 to automatically change the operating conditions to satisfy the operation command and control the operating indicators of the blast furnace 1.

[0058] The control unit 11 calculates how to control the equipment for inserting ore and coke in order to achieve the ore layer thickness ratio included in the target operating conditions. For example, if the blast furnace 1 is equipped with a bell-less device with a swivel chute, the control unit 11 may calculate the notch and rotations of the swivel chute. Also, for example, if the blast furnace 1 is equipped with a bell-type charging device, the control unit 11 may calculate at least one of the large bell opening degree, the large bell opening speed, and the movable armor stroke.

[0059] The control unit 11 may use the current ore layer thickness ratio measurement data when calculating how to control the apparatus for inserting ore and coke. The current ore layer thickness ratio may be measured, for example, by a distance meter (profile meter) installed inside the blast furnace 1.

[0060] The control unit 11 calculates how to control the ore and coke insertion apparatus in order to achieve the ore layer thickness ratio included in the target operating conditions, and then controls the ore and coke insertion apparatus to satisfy the calculated conditions.

[0061] The operation method of the blast furnace according to this embodiment will be explained with reference to the flowchart shown in Figure 2.

[0062] Step S101: The control unit 11 of the operating condition control device 10 acquires an operating command to bring the operating indicators showing the state of the blast furnace 1 within the target range. Triggered by the acquisition of the operating command, the control unit 11 starts processing the flowchart shown in Figure 2.

[0063] Step S102: When the control unit 11 receives an operation command, it calculates target operating conditions based on the operation indicator estimation model to bring the difference between the operation indicator after changing the operating conditions and the target operation indicator included in the operation command within a predetermined range.

[0064] Step S103: Once the control unit 11 calculates the target operating conditions, it automatically changes the current operating conditions to the target operating conditions. Alternatively, the control unit 11 may output the calculated target operating conditions to the output unit 13 and present them to the operator. This allows the operator, having grasped the target operating conditions, to change the current operating conditions to the target operating conditions and control the operating indicators of the blast furnace 1 to meet the operating command.

[0065] (Example 1) This section describes an example of the actual application of the blast furnace operation method according to this embodiment.

[0066] The operating method of a blast furnace according to this embodiment is described for a furnace with an internal volume of 4000 m³. 3 This method was applied to a bellless blast furnace of the same class. Using five years of historical operating data, an operational indicator estimation model was created using gas utilization rate and upper ventilation resistance index as operational indicators.

[0067] The gas utilization rate (represented as ηCO) was calculated by sampling the gas at the top of blast furnace 1, measuring the CO concentration and CO2 concentration, and using the value shown in equation (2) below.

number

[0068] Furthermore, the upper airflow resistance index (represented as K) was defined by the following equation (3), using the average value of pressure gauges installed in the same cross-section in the shaft section of blast furnace 1, the differential pressure ΔP between the pressure gauge installed at the top of the furnace, the height distance L between the two points, and the flow velocity U of the gas inside the furnace.

number

[0069] Regarding the heat load, it is desirable to consider both the amount of heat dissipation calculated from the blast furnace shell temperature for each part and the amount of heat removed by the cooling water based on the temperature difference between the inlet and outlet of the cooling water. However, since the majority of the heat removed from the blast furnace can be expressed as heat removed by the cooling water, in this embodiment, the heat load is defined by the following equation (4) based on the temperature difference between the inlet and outlet of the cooling water.

number

[0070] Here, the gas flow rate distribution inside the blast furnace can be approximated by estimating the temperature drop from the temperature at the top of the furnace, assuming that the gas temperature is constant at a certain height, based on the gas temperature measured at each location and the heat capacity of the solid obtained from the ore layer thickness ratio at each location. Alternatively, the flow velocity distribution inside the blast furnace can be obtained using flow meters such as Pitot tubes installed at each location. If the ventilation resistance in the ore layer and coke layer can be defined in advance, actions can be set to achieve the target.

[0071] The operational indicator estimation model, used to estimate these operational indicators, utilizes parameters such as oxygen enrichment rate, blast flow rate, ironmaking rate, pulverized coal ratio, heat flow ratio, sintered ore ratio, small coke lump ratio, coke strength, sintered ore strength, sintered ore moisture content, blast temperature, and blast furnace gas flow rate distribution as operational conditions. By considering so many operational conditions, the operational indicator estimation model can estimate operational indicators with high accuracy. In addition to the above, the operational indicator estimation model also uses the distribution of ore layer thickness ratios, calculated from the output of the furnace profile meter and measured at six points in the furnace radial direction. By using values ​​from multiple locations as the distribution of ore layer thickness ratios, the operational indicator estimation model can improve the accuracy of its estimation of operational indicators.

[0072] The ore layer thickness ratio may be used as is, or it may be used as a feature obtained by processing the measurement results. If used as a feature, for example, the mean values, gradients, and standard deviations of the central, intermediate, and outer regions may be used as features.

[0073] Inside the blast furnace, it takes approximately 8 hours for the charged raw materials to be discharged as molten iron from the bottom of the furnace. Therefore, all 8 parameter values ​​from 8 hours prior to the present were used for model creation. However, for the ore layer thickness ratio, which uses the output value of the profile meter, only one data point is available every 8 hours, so the hourly data was not used. For the ore layer thickness ratio, only the data for the ore layer thickness ratio at the time for which data is available was used to create the operational indicator estimation model.

[0074] As a specific model for estimating operational indicators, we adopted the linear model ElasticNet. However, the type of model used is not limited to ElasticNet. It is desirable to use any model with high accuracy, not just ElasticNet.

[0075] In this embodiment, only operational values ​​obtained every 8 hours, and only from points where operations were stable, were used as historical data to create the operational indicator estimation model. Therefore, approximately 5,000 training data points were used as 5 years of historical operational data.

[0076] Each parameter used in the operational indicator estimation model had 8 points per hour, resulting in a total of approximately 300 parameters. Using all parameters could lead to overfitting and multicollinearity, potentially preventing the creation of a proper model. Therefore, ElasticNet was subjected to L1 and L2 regularization to minimize these effects. Approximately 5000 measurement points were divided into an 8:2 time series. The ElasticNet model was trained using 80% of the data points, varying the parameters, and the operational indicator estimation model was created using the remaining 20% ​​of the data points with the most accurate parameters.

[0077] Figure 3 shows the results of estimating the upper ventilation resistance index using the operational indicator estimation model obtained after training. The vertical axis represents the estimated value of the upper ventilation resistance index. The horizontal axis represents the actual value of the upper ventilation resistance index. Circles without borders indicate the estimation results for 80% of the data used for training. Circles with borders indicate the estimation results for 20% of the data used for validation. Referring to Figure 3, it can be confirmed that the upper ventilation resistance index was estimated well in both the 80% of data used for training and the 20% of data used for validation.

[0078] Next, we will show the results of modifying operations in response to the operational instruction using the operational indicator estimation model described above. The operational instruction was to keep the gas utilization rate constant while lowering the upper ventilation resistance index.

[0079] In response to this operation order, the ore layer thickness ratio, which is an operating condition, was calculated and changed. Figure 4 shows the ore layer thickness ratio before and after the change. The changed ore layer thickness ratio is the ore layer thickness ratio calculated in response to the operation order. The horizontal axis represents the radial position within the blast furnace, and is shown by dividing the distance from the center of the blast furnace by the radius of the blast furnace body. Therefore, 0 on the horizontal axis means the center of the blast furnace, and 1.0 on the horizontal axis means the outer periphery of the blast furnace. The vertical axis shows the ore layer thickness ratio.

[0080] Ideally, a low ventilation resistance index is desirable to prevent problems such as open spaces. Therefore, an operational command that minimizes the ventilation resistance index would be acceptable. However, in this embodiment, in order to verify the accuracy of the operational indicator estimation model, an operational command was issued that lowered the ventilation resistance index to a certain extent.

[0081] Figures 5 and 6 show the results of changes in operational indicators when the ore layer thickness ratio is changed as shown in Figure 4. Figure 5 shows the results of changes in the upper ventilation resistance index. Figure 6 shows the results of changes in the gas utilization rate.

[0082] Referring to Figure 5, it can be seen that both actual and estimated values ​​show that the upper air permeability resistance index can be reduced by changing the ore layer thickness ratio.

[0083] Referring to Figure 6, it can be seen that, both in actual and estimated values, the gas utilization rate can be kept almost constant even when the ore layer thickness ratio is changed.

[0084] (Example 2) In this example, the internal volume is approximately 4000 m³. 3 We constructed an operational indicator prediction model using operational data from the past five years for the bellless blast furnace. The operational indicators targeted were gas utilization rate (ηCO), aeration resistance index (upper section), and heat load.

[0085] The gas utilization rate (ηCO) was calculated using equation (2) with the CO concentration and CO2 concentration of the gas sampled from the top of the blast furnace.

[0086] The airflow resistance index (upper part) was calculated using equation (3) based on the differential pressure ΔP between the average value of multiple pressure gauges installed in the blast furnace shaft and the furnace top pressure, the height L between measurement points, and the gas flow velocity U inside the furnace.

[0087] Ideally, the heat load should be determined by the sum of the heat dissipation based on the blast furnace shell temperature and the heat removed based on the temperature difference between the inlet and outlet of the cooling water. However, since the heat removed by the cooling water can represent the majority of the heat load, in this embodiment it is defined by the following equation (5) based on the heat removed by the cooling water.

number

[0088] The explanatory variables used in model construction can be classified into the following two groups. Group 1: Operating conditions such as airflow rate, enriched oxygen flow rate, air moisture content, furnace top pressure, air temperature, pulverized coal injection rate, pig iron output, coke ratio, pellet ratio, sintered ore ratio, lump ore ratio, coke strength index, and sintered strength index. Group 2: Ore layer thickness ratio distribution, heat flux ratio distribution, blast furnace gas flow rate distribution

[0089] In this example, for each operational indicator, a model was constructed that combined one of the operating conditions from Group 1 and one of the distributions from Group 2. Therefore, three types of predictive models were created for each operational indicator.

[0090] The ore layer thickness ratio distribution is the distribution of ore layer thickness ratios measured in the radial direction of the furnace, calculated from the output from the furnace profile meter.

[0091] The heat flow ratio is an index that represents the ratio of the heat capacity of solids and gases in the furnace, and strictly speaking, it is necessary to measure the coke-ore ratio, raw material descent rate, gas flow rate, gas temperature, and gas composition for each part. However, in this embodiment, a simplified method was adopted. Specifically, the heat flow ratio was calculated based on the specific heat of the solid raw materials in the entire blast furnace, the descent rate, gas flow rate, gas temperature, and gas composition, and the correlation with gas temperature was investigated. A linear relationship was found between the two. Therefore, the heat flow ratio distribution was estimated from the furnace gas temperature using this relationship.

[0092] While it is possible to directly measure the gas flow rate distribution within the blast furnace at each location, in this embodiment, it was simply estimated based on the temperature balance and oxygen balance within the blast furnace.

[0093] Distribution data for ore layer thickness ratio, heat flux ratio, and blast furnace gas flow rate were acquired at 10 points in the furnace radial direction. Principal component analysis (PCA) was applied, and it was confirmed that approximately 80% of the variance could be explained by the first to fifth principal components in each case. As an example, Figure 7 shows the distribution shapes of the first to fifth principal components of the blast furnace gas flow rate distribution. These principal component scores were used as explanatory variables, representing the respective distributions, in the model construction.

[0094] Since it takes approximately 8 hours for the charged raw materials to be discharged as molten iron in the blast furnace, 24-hour average values ​​were used for all parameters to reflect the stable state. A linear regression model was adopted for the prediction model, and a stepwise method based on the Akaike Information Criterion (AIC) was used for variable selection. Historical data was extracted on a daily basis, and only data from stable operation periods were extracted, with approximately 2000 points used as training data. The data was split 8:2 in a time series, with 80% used for training and 20% for validation.

[0095] As an example, Figure 8 shows the predicted results of the airflow resistance index when using the heat flux ratio distribution. Figure 9 shows the predicted results of the airflow resistance index when using the gas flow rate distribution.

[0096] Next, we will show the results of modifying operations in response to the operational command using the operational indicator estimation model described above. The operational command was the same as in Example 1, which instructed the system to maintain a constant gas utilization rate while lowering the upper ventilation resistance index.

[0097] Based on this operational instruction, the distribution of ore layer thickness ratio, which is an operating condition, was calculated, and the distribution of ore layer thickness ratio was modified. Figure 10 shows an example of the modified ore layer thickness ratio distribution calculated based on the operational instruction. The horizontal axis represents the radial position within the blast furnace, and is a dimensionless value obtained by dividing the distance from the blast furnace center by the furnace radius. Therefore, 0 on the horizontal axis represents the center of the blast furnace, and 1.0 represents the outer periphery of the blast furnace. The vertical axis represents the ore layer thickness ratio. The distribution calculated based on the operational instruction shows that it is necessary to increase the ore layer thickness ratio to the upper limit of the operating range in the radial position range of 0.5 to 0.7.

[0098] Figure 11A shows the results of the change in the air permeability resistance index when the ore layer thickness ratio is changed as shown in Figure 10. The vertical axis in Figure 11A represents the ratio when the air permeability resistance index value before the change in the ore layer thickness ratio distribution is set to 1.0. Referring to Figure 11A, it can be seen that the air permeability resistance index in the upper part can be reduced by 8% by changing the ore layer thickness ratio.

[0099] Furthermore, Figure 11B shows the gas utilization rate before and after the change. The vertical axis of Figure 11B represents the ratio of the gas utilization rate before the change in the distribution of ore layer thickness, with the value set to 1.0. Referring to Figure 11B, it can be seen that the gas utilization rate was maintained at a similar level before and after the change in the ore layer thickness ratio.

[0100] As described above, the blast furnace operation method according to this embodiment includes the steps of: obtaining an operation command to bring the operation indicators, which indicate the state of the blast furnace 1 when it is in operation, within a target range; and calculating target operating conditions, which are operating conditions that bring the difference between the operation indicators after changing the operating conditions and the target operating indicators included in the operation command within a predetermined range, based on an operation indicator estimation model that takes operating conditions, which are the conditions for operating the blast furnace 1, as input and output operating indicators. The operation command also includes target operating indicators for two or more operating indicators. The target operating indicator for one of the two or more operating indicators maximizes, minimizes, or sets the operating indicator to a predetermined value, and the target operating indicators for the other of the two or more operating indicators set the target operating indicator to a predetermined value, greater than or equal to a predetermined value, or less than or equal to a predetermined value. The operating conditions also include the distribution of ore layer thickness ratios. Thus, the blast furnace operation method according to this embodiment does not determine target operating conditions based on empirical rules, but rather calculates target operating conditions based on an operating indicator estimation model, and since the operation command includes target operating indicators for two or more operating indicators, the blast furnace operation method according to this embodiment can accurately calculate a desired ore bed thickness ratio that sets multiple operating indicators to predetermined targets.

[0101] Furthermore, in the blast furnace operation method according to this embodiment, the operating conditions further include at least one of the following: oxygen enrichment rate, blast flow rate, ironmaking rate, pulverized coal ratio, heat flow ratio, sintered ore ratio, small coke lump ratio, coke strength, sintered ore strength, sintered ore moisture content, blast temperature, and gas flow rate distribution inside the blast furnace. By considering so many operating conditions, the blast furnace operation method according to this embodiment can accurately calculate the desired operating conditions.

[0102] This disclosure is not limited to the embodiments described above. For example, multiple blocks described in the block diagram may be combined, or a single block may be divided. Instead of executing multiple steps described in the flowchart in chronological order as described, they may be executed in parallel or in a different order, depending on the processing capacity of the device performing each step, or as necessary. Other modifications are possible without departing from the spirit of this disclosure. [Explanation of Symbols]

[0103] 1 blast furnace 10. Operating Condition Control Device 11 Control Unit 12 Input section 13 Output section 14 Storage section 15 Communications Department

Claims

1. The steps include obtaining an operational command to bring the operational indicators showing the state of the blast furnace while it is in operation within a target range, A step of calculating target operating conditions, which are the operating conditions after changing the operating conditions, and which are the operating conditions included in the operation command, based on an operation indicator estimation model that takes operating conditions, which are the conditions for operating the blast furnace, as input and the operation indicator as output, so that the difference between the operation indicator after changing the operating conditions and the target operation indicator included in the operation command is within a predetermined range. Includes, The aforementioned operational command includes the aforementioned target operational indicators for two or more of the aforementioned operational indicators, The target operating indicator for one of the two or more operating indicators is to maximize, minimize, or set the operating indicator to a predetermined value. The target operating indicator for the other operating indicator among the two or more operating indicators is such that the target operating indicator is set to a predetermined value, greater than or equal to a predetermined value, or less than or equal to a predetermined value. The aforementioned operating conditions include the distribution of ore layer thickness ratios, and the method of operating a blast furnace.

2. A method for operating a blast furnace according to claim 1, further comprising the steps of presenting the calculated target operating conditions to an operator, or automatically changing the operating conditions based on the calculated target operating conditions.

3. The method for operating a blast furnace according to claim 1, wherein the operating indicators include at least one of the aeration resistance index, gas utilization rate, heat load, and solution loss carbon amount.

4. The method for operating a blast furnace according to claim 1, wherein the operating conditions further include at least one of oxygen enrichment rate, blast flow rate, iron production rate, pulverized coal ratio, heat flow ratio, sintered ore ratio, coke lump ratio, coke strength, sintered ore strength, sintered ore moisture content, blast temperature, and gas flow rate distribution inside the blast furnace.

5. The method for operating a blast furnace according to claim 4, wherein the distribution of the ore layer thickness ratio, the distribution of the gas flow rate inside the blast furnace, and the distribution of the heat flow ratio have two or more values ​​in the radial direction or cross-sectional direction inside the blast furnace, or are feature quantities that indicate the distribution in the radial direction or cross-sectional direction inside the blast furnace calculated from the two or more values.

6. The method for operating a blast furnace according to claim 5, wherein the feature quantities are principal component scores obtained by dimensionality reduction of the radial or cross-sectional distribution inside the blast furnace using principal component analysis (PCA).

7. The method for operating a blast furnace according to claim 6, characterized in that the feature quantities used are the first to Nth principal components (where N is an integer including 5) whose cumulative contribution rate is 70% or more and 90% or less.

8. The method for operating a blast furnace according to claim 1, wherein the operating indicator estimation model is a model constructed based on the past operating performance of the blast furnace.

Images