EP4674991A1

EP4674991A1 - Method and device for detecting air leak in sintering machine pallet

Info

Publication number: EP4674991A1
Application number: EP24806844.7A
Authority: EP
Inventors: Yoshiki Kitamura
Original assignee: JFE Steel Corp
Current assignee: JFE Steel Corp
Priority date: 2023-05-18
Filing date: 2024-03-07
Publication date: 2026-01-07
Also published as: JP7571919B1; JPWO2024236889A1

Abstract

An air leakage detection method for a sintering machine pallet according to the present invention includes: a measuring step of measuring, by using an oxygen concentration meter installed in a wind leg immediately below a plurality of circulating pallets, an oxygen concentration in the wind leg for each circulation cycle of each of the plurality of pallets when each of the plurality of pallets passes immediately above the wind leg; a predicting step of calculating a predicted value of the oxygen concentration by inputting information at a time of processing to a prediction model, the prediction model using the information indicating an operation state of the Dwight-Lloyd type sintering machine and a blending state of the sintering raw material as an input variable and the predicted value of the oxygen concentration as an output variable; and a detecting step of detecting a pallet having air leakage using a subtracted value obtained by subtracting the predicted value of the oxygen concentration calculated in the predicting step from the oxygen concentration for each circulation cycle of each of the plurality of pallets measured in the measuring step.

Description

Field

The present invention relates to an air leakage detection method and an air leakage detection device, for a sintering machine pallet, that detect air leakage in pallets of a Dwight-Lloyd type sintering machine for firing a sintering raw material.

Background

A Dwight-Lloyd type sintering machine fires a sintering raw material to produce a sintered ore. Specifically, in the Dwight-Lloyd type sintering machine, the sintering raw material is deposited on a circulating pallet, the surface of the raw material is ignited, and then air is sucked downward from the top of the pallet by a wind box. As a result, the sintering raw material is fired to produce the sintered ore. The air sucked by the wind box is discharged to a main duct via a wind leg.
When air leakage occurs in an air discharge system in the Dwight-Lloyd type sintering machine having the above configuration, unnecessary air suction is induced from an air-leaking part, and an air volume required for firing the sintering raw material is reduced. As a result, a progress of firing of the sintering raw material is adversely affected, and the productivity of the sintered ore decreases. Therefore, in order to maintain the productivity of the sintered ore, it is important to take measures such as repair by detecting an air-leaking part at an early stage. In view of the above background, a technique for detecting an air-leaking part in the Dwight-Lloyd type sintering machine has been proposed.
Specifically, Patent Literature 1 describes a technique of measuring an oxygen concentration in a combustion gas immediately below a circulating pallet by a laser oxygen concentration meter, identifying a pallet, and specifying an air-leaking part in the pallet from a change in the oxygen concentration as the pallet is circulated. Still more, Patent Literature 2 describes a technique of specifying an air-leaking part in a pallet based on an oxygen concentration distribution in each pallet and variation in the oxygen concentration distribution in each circulation cycle, inputting the air-leaking part to a history management unit as an operation history, and specifying a pallet that needs to be repaired based on the operation history.

Citation List

Patent Literature

Patent Literature 1: JP 2012-36477 A
Patent Literature 2: JP 2014-122380 A

Summary

Technical Problem

However, in techniques described in Patent Literature 1 and Patent Literature 2, it is not possible to distinguish variation in oxygen concentration due to variation in an operation state of the Dwight-Lloyd type sintering machine or a blending state of a sintering raw material from variation in oxygen concentration due to air leakage. Therefore, according to the techniques described in Patent Literature 1 and Patent Literature 2, a change in the oxygen concentration may be determined to be caused by air leakage, although the oxygen concentration has changed due to variation in the operation state of the Dwight-Lloyd type sintering machine or the blending state of the sintering raw material, and there is a possibility that the air leakage is erroneously detected.
The present invention has been made to solve the above problem, and an object thereof is to provide an air leakage detection method and an air leakage detection device, for a sintering machine pallet, capable of accurately detecting air leakage in the pallet.

Solution to Problem

A method for detecting air leakage in a sintering machine pallet of a Dwight-Lloyd type sintering machine for firing a sintering raw material according to the present invention includes: a measuring step of measuring, by using an oxygen concentration meter installed in a wind leg immediately below a plurality of circulating pallets, an oxygen concentration in the wind leg for each circulation cycle of each of the plurality of pallets when each of the plurality of pallets passes immediately above the wind leg; a predicting step of calculating a predicted value of the oxygen concentration by inputting information at a time of processing to a prediction model, the prediction model using the information indicating an operation state of the Dwight-Lloyd type sintering machine and a blending state of the sintering raw material as an input variable and the predicted value of the oxygen concentration as an output variable; and a detecting step of detecting a pallet having air leakage using a subtracted value obtained by subtracting the predicted value of the oxygen concentration calculated in the predicting step from the oxygen concentration for each circulation cycle of each of the plurality of pallets measured in the measuring step.
The prediction model may be a machine learning model trained by machine learning using a set of an actual value of the information indicating the operation state of the Dwight-Lloyd type sintering machine and the blending state of the sintering raw material and an actual value of the oxygen concentration as teaching data.
The information may include information on a coagulant basic unit, a quicklime blending ratio, an air volume of a main exhauster, a pressure in a main duct, and a temperature in a wind box.
A device for detecting air leakage in a sintering machine pallet of a Dwight-Lloyd type sintering machine for firing a sintering raw material according to the present invention includes: an oxygen concentration meter configured to measure, by using an oxygen concentration meter installed in a wind leg immediately below a plurality of circulating pallets, an oxygen concentration in the wind leg for each circulation cycle of each of the plurality of pallets when each of the plurality of pallets passes immediately above the wind leg; and a control device configured to calculate a predicted value of the oxygen concentration by inputting information at a time of processing to a prediction model, the prediction model using the information indicating an operation state of the Dwight-Lloyd type sintering machine and a blending state of the sintering raw material as an input variable and the predicted value of the oxygen concentration as an output variable, and detect a pallet having air leakage using a subtracted value obtained by subtracting the calculated predicted value of the oxygen concentration from the measured oxygen concentration for each circulation cycle of each of the plurality of pallets.

Advantageous Effects of Invention

According to an air leakage detection method and an air leakage detection device for a sintering machine pallet of the present invention, air leakage in a pallet is accurately detectable.

Brief Description of Drawings

FIG. 1 is a schematic diagram illustrating a configuration of a sintering machine to which an air leakage detection method and an air leakage detection device for a sintering machine pallet are applied according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating an example of a change in an actual value of oxygen concentration measured by an oxygen concentration meter as a pallet is circulated.
FIG. 3 is a conceptual diagram illustrating components included in an oxygen concentration measured by the oxygen concentration meter.
FIG. 4 is a diagram illustrating an example of a change in prediction accuracy of an oxygen concentration prediction model as the number of explanatory variables changes.
FIG. 5 is a diagram illustrating an example of a change in the oxygen concentration after removal of an operation-affected component as the pallet is circulated.
FIG. 6 is a flowchart illustrating a flow of an air leakage detection process according to the embodiment of the present invention.

Description of Embodiments

Hereinafter, an air leakage detection method and an air leakage detection device for a sintering machine pallet according to an embodiment of the present invention will be described with reference to the drawings.

[Configuration]

First, a configuration of a sintering machine to which the air leakage detection method and the air leakage detection device for a sintering machine pallet, according to the embodiment of the present invention, are applied will be described with reference to FIG. 1.
FIG. 1 is a schematic diagram illustrating the configuration of the sintering machine to which the air leakage detection method and the air leakage detection device for a sintering machine pallet, according to the embodiment of the present invention, are applied. As illustrated in FIG. 1, the sintering machine to which the air leakage detection method and the air leakage detection device for a sintering machine pallet, according to the embodiment of the present invention, are applied is configured by a Dwight-Lloyd type sintering machine 1. As illustrated in FIG. 1, the Dwight-Lloyd type sintering machine 1 includes a plurality of pallets 2, a charging device 3 (3a and 3b), an ignition furnace 4, a plurality of wind boxes 5 and wind legs 6, a main exhauster 7, and a crusher 8.
The plurality of pallets 2 are devices that are endlessly connected between two sprockets 11a and 11b, and circulate in an arrow direction. The charging device 3 is a device for depositing a sintering raw material in layers on the pallets 2. The ignition furnace 4 is a device that ignites the sintering raw material deposited on the pallets 2. The plurality of wind boxes 5 are devices that suck air from the top to bottom of the pallet 2 through the sintering raw material deposited and continuously arranged on the plurality of pallets 2 in a circulating direction. The main exhauster 7 is a device that discharges air sucked by the plurality of wind boxes 5 to the outside, as exhaust gas, via the wind legs 6 and a main duct 12 connected to a lower part of the respective wind boxes 5. The crusher 8 is a device that crushes a sintered ore generated by firing the sintering raw material on the pallet 2.
Still more, the Dwight-Lloyd type sintering machine 1 includes an oxygen concentration meter 21 and a control device 22 as a control system. The oxygen concentration meter 21 is configured with an oxygen concentration meter such as a zirconia type oxygen densitometer, and is disposed at a lower part of an arbitrary wind leg 6a among the plurality of wind legs 6. The oxygen concentration meter 21 measures the oxygen concentration in air sucked from a pallet 2a passing immediately above a wind box 5a, and inputs an electric signal indicating the oxygen concentration measured to the control device 22. The control device 22 is configured with an information processing apparatus such as a computer, and controls the operation of the Dwight-Lloyd type sintering machine 1. Further, in the present embodiment, the control device 22 detects air leakage in the pallets 2 based on the oxygen concentration measured by the oxygen concentration meter 21. When the air leakage is detected, the control device 22 notifies information or the like on a position of the pallet 2 having air leakage.
In this Dwight-Lloyd type sintering machine 1, first, the sintering raw material including iron ore powder, powder collected in a steel mill, sintered ore undersize powder, CaO-containing raw materials (CaO-based fluxes) such as limestone and dolomite, a granulation aid such as quicklime, and a coagulant such as coke breeze and anthracite is stored in the charging device 3. Next, the sintering raw material is charged from the charging device 3 onto each pallet 2 to form a raw material layer. Next, the coagulant on an upper surface of the raw material layer is ignited by the ignition furnace 4, and air is sucked from each wind box 5 to sequentially burn the coagulant from an upper layer to a lower layer of the raw material layer. The sintering raw material is heated and melted by combustion heat generated at this time. As a result, a sintered layer (sintered cake) is generated. Finally, the sintered layer formed on each pallet 2 is crushed by the crusher 8 and collected as the sintered ore.

[Air leakage detection process]

Next, the air leakage detection method in the pallets 2 of the Dwight-Lloyd type sintering machine 1 illustrated in FIG. 1 will be described with reference to FIGS. 2 to 6.
FIG. 2 is a diagram illustrating an example of a change in an actual value of oxygen concentration measured by the oxygen concentration meter 21 as the pallet 2 is circulated. As illustrated in FIG. 2, it can be confirmed that there is periodic variation in the actual value of oxygen concentration (actual oxygen concentration) measured by the oxygen concentration meter 21. However, even in the same pallet 2 (pallet position), an absolute value of the actual oxygen concentration varies (mean absolute error (MAE) with respect to reference value: 0.80). Therefore, it is difficult to accurately detect the air leakage in the pallet 2 from the change in actual oxygen concentration.
Thus, the inventor of the present invention has considered that a reason that the actual oxygen concentration varies is that, as illustrated in FIG. 3, the oxygen concentration includes an operation-affected component that is affected by variation in an operation state (operation conditions, exhaust gas state, and the like) of the Dwight-Lloyd type sintering machine 1 and a blending state of the sintering raw material, and an air-leakage-affected component that is affected by air leakage in the pallet 2. Then, it is considered that the air leakage in the pallet 2 can be accurately detected by detecting the air leakage in the pallet 2 based on the oxygen concentration after removing the operation-affected component.
Specifically, first, the inventor has built a prediction model for predicting the oxygen concentration measured by the oxygen concentration meter 21. More specifically, multiple regression analysis is performed, using actual values of various parameters related to the operation state of the Dwight-Lloyd type sintering machine 1 as explanatory variables, and actual values of the oxygen concentration measured by the oxygen concentration meter 21 as objective variables. Then, an accuracy of an oxygen concentration prediction model obtained by the multiple regression analysis is evaluated. The oxygen concentration prediction model uses parameters related to the operation state of the Dwight-Lloyd type sintering machine 1 as input variables, and predicted values of the oxygen concentration at an installation position of the oxygen concentration meter 21 as output variables.

As a result, as illustrated in FIG. 4, MAE of the predicted values of the oxygen concentration (output of the oxygen concentration prediction model) and the actual values is minimized when the explanatory variables are five types illustrated in Table 1 below. It has been confirmed that the oxygen concentration at the installation position of the oxygen concentration meter 21 can be accurately predicted. Further, as illustrated in Table 1, it has been confirmed that the importance as the explanatory variable increases in the order of a coagulant basic unit, a quicklime blending ratio, an air volume of the main exhauster 7, a pressure in the main duct 12, and a temperature in the wind box 5a.

Table 1

Order of importance	Explanatory variable	Coefficient	Importance
1	Wind box temperature	-0.092	0.518
2	Main duct pressure	-0.207	0.319
3	Air volume of exhauster	-0.007	0.088
4	Quicklime blending ratio	-0.304	0.051
5	Coagulant basic unit	-0.066	0.024

Next, an explanatory variable of a prediction target is input to the above oxygen concentration prediction model, and a predicted value of the oxygen concentration output from the oxygen concentration prediction model is subtracted from the actual value of the oxygen concentration (FIG. 2) measured by the oxygen concentration meter 21 to evaluate a change in a subtracted value as the pallet 2 is circulated. As a result, as illustrated in FIG. 5, MAE with respect to the reference value is 0.69, and it is confirmed that variation in oxygen concentration is reduced. Therefore, according to the change in oxygen concentration illustrated in FIG. 5, the air leakage in the pallet 2 can be accurately detected.
In the above description, the oxygen concentration prediction model is constructed using the multiple regression analysis. However, the method of constructing the oxygen concentration prediction model is not limited to the multiple regression analysis, and various machine learning methods can be used. Here, examples of the machine learning method include a generally used neural network (including deep learning and convolutional neural network), decision tree learning, random forest, and support vector regression. In this case, machine learning is performed using, as teaching data, a set of actual values of various parameters related to the operation state of the Dwight-Lloyd type sintering machine 1 and actual values of the oxygen concentration measured by the oxygen concentration meter 21. As a result, it is possible to construct, as the oxygen concentration prediction model, a machine learning model that uses parameters related to the operation state of the Dwight-Lloyd type sintering machine 1 as input variables and predicted values of the oxygen concentration at the installation position of the oxygen concentration meter 21 as output variables.
Hereinafter, a flow of the air leakage detection process according to the embodiment of the present invention based on the above idea will be described with reference to FIG. 6.
FIG. 6 is a flowchart illustrating the flow of the air leakage detection process according to the embodiment of the present invention. The flowchart illustrated in FIG. 6 starts at the timing when the operation of the Dwight-Lloyd type sintering machine 1 is started, and the air leakage detection process proceeds to Step S1.
In Step S1, the control device 22 acquires information on the oxygen concentration for each circulation cycle of each pallet 2 based on the electric signal input from the oxygen concentration meter 21. Then, Step S1 is completed, and the air leakage detection process proceeds to Step S2.
In Step S2, the control device 22 acquires tracking information indicating a position of the pallet 2 on a circulation orbit for identifying the pallet 2 that has passed immediately above the wind leg 6a where the oxygen concentration meter 21 is installed. Specifically, the control device 22 generates the tracking information based on an initial position on the circulation orbit of the pallet 2, a circulation speed, elapsed time from the start of circulation, and a distance of one circulation cycle. Then, Step S2 is completed, and the air leakage detection process proceeds to Step S3.
In Step S3, the control device 22 associates the oxygen concentration for each circulation cycle of each pallet 2 acquired in Step S1 with the tracking information of the pallet 2 acquired in Step S2, thereby generating data indicating a relationship between the position of the pallet 2 and the oxygen concentration as illustrated in FIG. 2. Then, Step S3 is completed, and the air leakage detection process proceeds to Step S4.
In Step S4, the control device 22 acquires input data of the oxygen concentration prediction model when each pallet 2 passes immediately above the wind leg 6a where the oxygen concentration meter 21 is installed, and inputs the input data acquired to the oxygen concentration prediction model. Then, the control device 22 subtracts the predicted value of the oxygen concentration that is output from the oxygen concentration prediction model from the oxygen concentration for each circulation cycle of each pallet 2, thereby removing a variation amount of the oxygen concentration, due to variation in the operation state of the Dwight-Lloyd type sintering machine 1 and the blending state of the sintering raw material, from the oxygen concentration for each circulation cycle of each pallet 2. As a result, as illustrated in FIG. 5, it is possible to generate data indicating a relationship between the position of the pallet 2 and a change in oxygen concentration after removal of the operational variation component. The control device 22 may perform the association process after removing the variation amount of the oxygen concentration, due to variation in the operating condition of the Dwight-Lloyd type sintering machine 1 and the blending state of the sintering raw material, without executing Step S3. Then, Step S4 is completed, and the air leakage detection process proceeds to Step S5.
In Step S5, the control device 22 detects air leakage in the pallet 2 using the data indicating the relationship between the position of the pallet 2 and the change in oxygen concentration after removal of the operational variation component. When the air leakage is detected, the control device 22 notifies information or the like on the position of the pallet 2 having air leakage. Upon receiving the notification from the control device 22, an operator performs repair work on the pallet 2 having air leakage. Then, Step S5 is completed, and a series of air leakage detection processes ends.
Although the embodiment to which the invention by the inventor is applied has been described above, the present invention is not limited to the description and drawings that form part of the disclosure of the present invention according to the present embodiment. In other words, other embodiments, examples, operation techniques, and the like made by those skilled in the art based on the present embodiment are all included in the scope of the present invention.

Industrial Applicability

According to the present invention, it is possible to provide an air leakage detection method and an air leakage detection device, for a sintering machine pallet, that are capable of accurately detecting air leakage in a pallet.

Reference Signs List

1 DWIGHT-LLOYD TYPE SINTERING MACHINE
2, 2a PALLET
3, 3a, 3b CHARGING DEVICE
4 IGNITION FURNACE
5, 5a WIND BOX
6, 6a WIND LEG
7 MAIN EXHAUSTER
8 CRUSHER
11a, 11b SPROCKET
12 MAIN DUCT
21 OXYGEN CONCENTRATION METER
22 CONTROL DEVICE

Claims

A method for detecting air leakage in a sintering machine pallet of a Dwight-Lloyd type sintering machine for firing a sintering raw material, the method comprising:
a measuring step of measuring, by using an oxygen concentration meter installed in a wind leg immediately below a plurality of circulating pallets, an oxygen concentration in the wind leg for each circulation cycle of each of the plurality of pallets when each of the plurality of pallets passes immediately above the wind leg;

a predicting step of calculating a predicted value of the oxygen concentration by inputting information at a time of processing to a prediction model, the prediction model using the information indicating an operation state of the Dwight-Lloyd type sintering machine and a blending state of the sintering raw material as an input variable and the predicted value of the oxygen concentration as an output variable; and

a detecting step of detecting a pallet having air leakage using a subtracted value obtained by subtracting the predicted value of the oxygen concentration calculated in the predicting step from the oxygen concentration for each circulation cycle of each of the plurality of pallets measured in the measuring step.
The method for detecting air leakage in a sintering machine pallet according to claim 1, wherein the prediction model is a machine learning model trained by machine learning using a set of an actual value of the information indicating the operation state of the Dwight-Lloyd type sintering machine and the blending state of the sintering raw material and an actual value of the oxygen concentration as teaching data.
The method for detecting air leakage in a sintering machine pallet according to claim 1 or 2, wherein the information includes information on a coagulant basic unit, a quicklime blending ratio, an air volume of a main exhauster, a pressure in a main duct, and a temperature in a wind box.
A device for detecting air leakage in a sintering machine pallet of a Dwight-Lloyd type sintering machine for firing a sintering raw material, the device comprising:
an oxygen concentration meter configured to measure, by using an oxygen concentration meter installed in a wind leg immediately below a plurality of circulating pallets, an oxygen concentration in the wind leg for each circulation cycle of each of the plurality of pallets when each of the plurality of pallets passes immediately above the wind leg; and

a control device configured to
calculate a predicted value of the oxygen concentration by inputting information at a time of processing to a prediction model, the prediction model using the information indicating an operation state of the Dwight-Lloyd type sintering machine and a blending state of the sintering raw material as an input variable and the predicted value of the oxygen concentration as an output variable, and

detect a pallet having air leakage using a subtracted value obtained by subtracting the calculated predicted value of the oxygen concentration from the measured oxygen concentration for each circulation cycle of each of the plurality of pallets.