WO2019124768A1

WO2019124768A1 - Sintering operation control apparatus and method therefor

Info

Publication number: WO2019124768A1
Application number: PCT/KR2018/014286
Authority: WO
Inventors: 손상한; 전지원; 나지훈; 정인현
Original assignee: 주식회사 포스코
Priority date: 2017-12-18
Filing date: 2018-11-20
Publication date: 2019-06-27
Also published as: KR20190072997A

Abstract

A sintering operation control apparatus may comprise: a database for storing a prediction model for predicting at least one manipulated value related to a sintering operation through a neural network algorithm; an operation data collection unit for acquiring a particle size distribution of sintered ore, discharged from a sintering machine, from an image obtained by capturing the sintered ore by using at least one camera, and collecting operation data including the particle size distribution in relation to an operation state of the sintering operation; a prediction unit for acquiring the at least one manipulated value by using the operation data as input data of the prediction model; and a control unit for automatically controlling sintering equipment according to the at least one manipulated value.

Description

Sintering operation control device and method thereof

An embodiment relates to a sintering operation control device and a method thereof.

The sintering process is a process in which a mixed raw material in which various kinds of iron ores, subsidiary raw materials such as limestone, and a binder such as coal are mixed is charged into a movable type sintering machine and then the partial melting of the iron ore is performed by the heat of combustion of the binder to be. The compacted sintered ores through the sintering process are used as raw materials for the blast furnace.

The ratio of the binder contained in the blending raw material, the height of the blending raw material on the sintering machine bed, the charging density of the blending raw material, and the moving speed of the sintering machine are very important factors for determining the productivity of the sintering operation and the quality of the sintered ores. Therefore, in the sintering step, the amount of the binder contained in the blending raw material, the height of the bed layer and the charging density of the blending raw material when the blending raw material is charged into the sintering machine, and the moving speed of the sintering machine It needs to be adjusted appropriately according to quality conditions.

On the other hand, the quality of the sinter ore produced during the sintering operation and the productivity of the sintering operation can be measured about once to three times per day due to various obstacles. When the operating factors of the sintering operation are controlled according to the quality of the sintered ore and the productivity of the sintering operation intermittently measured, there is a problem that it is difficult to efficiently manage the sintering operation due to the representative limit of the data.

An object of the present invention is to provide a sintering operation control apparatus and method capable of automatically controlling a sintering operation by judging a current sintering operation state in real time.

According to an aspect of the present invention, there is provided a sintering operation control apparatus including a database for storing a prediction model for predicting at least one operation value associated with sintering operation through a neural network algorithm, An operation data collecting unit for acquiring a particle size distribution of the sintered ores from the image of the sintered ores emitted from the sintering unit and collecting the operation data including the particle size distribution in relation to the operation state of the sintering operation, And a control unit for automatically controlling the sintering equipment in accordance with the at least one operation value.

The at least one operating value may include a binding material ratio in the blending raw material charged into the sintering machine, a blending raw material height on the bed of the sintering machine, a blending raw material loading density of the sintering machine, and a moving speed of the sintering machine .

The control unit may control an amount of the binder to be supplied from the storage bin to the mixer for raw material mixture, based on the combination ratio in the blended material among the at least one operation value.

The control unit may control the mixing material feed amount of the mixing material storage hopper based on the mixing material height on the bed of the sintering machine among the at least one operating value.

The control unit may control the pressing member on the sintering unit based on the mixing material loading density of the sintering unit among the at least one operating value. The control unit may control the conveyor that moves the sintering machine based on the moving speed of the sintering machine among the at least one operation value.

The neural network algorithm may be composed of a convolutional neural network.

The sintering operation control apparatus may further include a learning unit configured to learn the operation data collected through the operation data collecting unit with learning data for a predetermined period of time to construct the prediction model.

The operating data includes at least one of a raw material mixture ratio, a blending raw material component, a coupling ratio in a blending raw material, a binder particle size, an ignition temperature and a gas pressure in an ignition, a moisture content of the blending raw material, a dust generating amount, The flow rate of the sintering machine, the flow speed of the sintering machine, the charging density, the air permeability, the exhaust gas temperature, the oxygen concentration, the negative pressure and flow rate of the hood, the sintered light component, the sintered light particle size, the sintered light intensity, And may include at least one.

The sintering operation control method according to an embodiment of the present invention includes the steps of obtaining a particle size distribution of the sintered ores from an image of the sintered ores discharged from a sintering machine using at least one camera, Collecting operational data including the particle size distribution, using the operational data as input data of a prediction model predicting at least one operational value associated with the sintering operation through a neural network algorithm, Obtaining an operating value, and automatically controlling the sintering facility in accordance with the at least one operating value.

Wherein the at least one operating value includes at least one of a binding material ratio in the blended material charged into the sintering machine, a blending material height on the bed of the sintering machine, a mixing material loading density of the sintering machine, and a balancing speed of the sintering machine can do.

The controlling may include controlling an amount of the binder to be supplied from the storage bin to the mixer for raw material, based on the binder ratio in the blend material among the at least one operation value.

The controlling step may include controlling the feed amount of the compounding raw material storage hopper based on the height of the compounding raw material on the bed of the sintering machine among the at least one operation value.

The controlling may include controlling the pressing member on the sintering machine based on the mixing material loading density of the sintering machine among the at least one operating value.

The controlling may include controlling a conveyor that moves the sintering machine, based on the moving speed of the sintering machine among the at least one manipulated value.

In the sintering operation control method, the neural network algorithm may be composed of a convolution neural network.

The sintering operation control method may further include a step of constructing the prediction model by learning the operation data with training data for a predetermined period.

According to the embodiment, the sintering equipment can be controlled in real time in accordance with the current operating conditions, and the efficiency of the sintering operation can be improved and the quality of the sintered ores can be stably maintained.

Fig. 1 shows an example of a sintering facility.

2 is a schematic view of a sintering operation control apparatus according to an embodiment of the present invention.

3 is a diagram for explaining a method of verifying a prediction model according to an embodiment of the present invention.

4 is a schematic view illustrating a sintering operation control method according to an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art to which the present invention pertains. The present invention may be embodied in many different forms and is not limited to the embodiments described herein.

In order to clearly illustrate the embodiments of the present invention, portions that are not related to the description are omitted, and the same or similar components are denoted by the same reference numerals throughout the specification.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between .

Hereinafter, the sintering operation control device and the method thereof will be described in detail with reference to necessary drawings.

Figure 1 schematically shows a sintering facility.

1, the sintering equipment includes a mixing material storage hopper 10, an ignition furnace 20, a plurality of sintering machines 30, a conveyor 40, a windbox 50, a duct 61, a dust collector 62 A duct blower 63, a hood 70, and a heat source storage unit 80.

The raw materials such as iron ores, quartz, serpentine, and limestone as well as binders such as anthracite coal and coke are transported from a storage bin (not shown) to a mixer (not shown), mixed with water, granulation, and is charged into the compounding material storage hopper 10. The blending raw materials for sintering stored in the blending raw material storage hopper 10 are supplied to the sintering machine 30 at a predetermined ratio for sintering operation.

Each of the sintering machines 30 is of a moving type and receives the raw material mixture from the mixing raw material storage hopper 10 and sequentially moves horizontally in the sintering process progressing direction. Each sintering machine 30 moves to the light distribution section through horizontal movement, and discharges the sintered light that has been sintered to the light distribution section.

The ignition furnace 20 performs a function of burning the binder contained in the raw material layer by spraying a flame with a blending raw material on the bed of the sintering machine 30 (hereinafter referred to as a raw material layer). The sintering furnace 20 is located at one side of the mixing material storage hopper 10 and the sintering machine 30 supplied with the mixing material for sintering by the mixing material storage hopper 10 The flame is injected into the raw material layer of the sintering machine 30. The ignition furnace 20 may use various means capable of injecting a flame into a raw material layer such as a gas burner.

A hood 70 for supplying a gas for supplying a heat source to the sintering machine 30 is disposed at one side of the ignition furnace 20 and the hood 70 is connected to a heat source storage part 80 for storing a gas for supplying a heat source .

The conveyor 40 is extended in the sintering process progressing direction, and moves a plurality of sintering machines 30 sequentially disposed in the sintering process progressing direction in the sintering process progressing direction.

A plurality of windboxes 50 are disposed on the movement path of the sintering machine 30 to provide a suction force to each of the sintering machines 30 so that the outside air sucked by the sintering machine 30, Gas, and flame ignited by the ignition furnace 20 to move toward the lower side of the sintering machine 30.

The dust sucked by the plurality of wind boxes 50 is removed by the duct 61, the dust collector 62, and the duct blower 63.

In the sintering plant having the above-described structure, when the flame is sprayed on the upper part of the raw material layer charged in the sintering machine 30 by the ignition furnace 30, the heat due to the flame, the outside air sucked into the sintering machine 30, The upper layer, that is, the surface layer of the raw material layer is ignited. As a result, the flame temperature is raised to 1300 ° C. to 1400 ° C., whereby the subsidiary raw materials such as limestone and iron ores form a low melting point compound and are partially melted to proceed the sintering reaction of iron ores.

The sintering machine 30 in which the surface layer of the raw material layer is ignited by the ignition furnace 20 is conveyed in one direction by the conveyor 40 and the ignited raw material layer is connected to the cold air by the movement of the sintering machine 30 And cooled. A wind box 50 is provided below the sintering machine 30, and the wind box 50 sucks the heat of the ignited raw material layer downward. At this time, the heat of the raw material layer gradually descends from the upper layer portion to the lower layer portion due to the suction force of the wind box 50, whereby the entire raw material layer is fired, and the fired raw material layer moves along the sintering machine 30 Cooled and finally made into sintered ores.

In this sintering process, the bonding material ratio included in the blending raw material, the height of the raw material layer on the sintering machine 30 bed, the loading density of the raw material layer, and the moving speed of the sintering machine 30, It is a very important factor to decide.

Therefore, in the embodiment of the present invention, the combination ratio, the height of the raw material layer on the bed of the sintering machine 30, the loading density of the raw material layer, and the conveyance speed of the sintering machine 30 And controls the sintering equipment on the basis thereof.

2 is a schematic view of a blast furnace operation control apparatus according to an embodiment of the present invention. 3 is a diagram for explaining a method of verifying a prediction model according to an embodiment of the present invention.

2, a sintering operation control apparatus 100 according to an embodiment of the present invention includes a operation data collecting unit 110, a user input unit 120, a operation data database 130, a learning unit 140, A database 150, a predicting unit 160, and a sintering operation control unit 170.

The operation data collecting unit 110 can collect a plurality of operation data related to the operating state of the sintering operation in real time.

Table 1 below shows examples of the operation data collected by the operation data collecting unit 110.

Table 1. Examples of operating data

Referring to Table 1, the operation data may include raw material mixture data, raw material mixture data, operation data of the ignition system 20, operation data of the sintering machine 30, sintering light data, amount of dust generated, and the like. The raw material blend data may include information related to blending raw materials such as raw material blending ratio (ore kind, sub-raw material type, particle size), blended raw material ingredient, binding raw material ratio in blended raw material, binder particle size and the like. In addition, the raw material mixture data may include information related to the raw material mixture such as the primary and secondary moisture contents of the raw material mixture. The operation data of the ignition circuit 20 may also include information related to the operation of the ignition circuit 20, such as the ignition temperature of the ignition circuit 20, the gas pressure of the ignition circuit 20, and the like. The operating data of the sintering machine 30 is used to calculate the feed rate and loading ratio of the raw material to the sintering machine 30, the post-layering (the height of the compounding material on the bed of the sintering machine 30), the sintering advancing speed , Air permeability, exhaust gas temperature, negative pressure and flow rate (or air volume) of hood 70, oxygen concentration, and the like. The sintered-orbital data also includes information related to the sintered ores produced by the sintering operation such as the components (sintered ores) of the sintered ores (FeO content, basicity etc.), particle size (or particle size distribution) and strength .

The operation data collecting unit 110 includes at least one sensor (not shown), and can automatically acquire operation data using the at least one sensor. For example, the operation data collecting unit 110 may acquire various sensed values by using a temperature sensor, a pressure sensor, a gas sensor, and the like, and obtain operation data related to the sintering operation therefrom.

The operation data collecting unit 110 includes at least one camera (not shown), and can automatically acquire operation data using the at least one camera (not shown). For example, the operation data collecting unit 110 photographs the compounding materials accumulated on the bed of the sintering machine 30 in real time using a camera, and automatically acquires the layering information through image analysis of the image It is possible. In addition, for example, the operation data collecting unit 110 photographs the mixed raw materials charged into the sintering machine 30 in real time through at least one camera, and implements the filling density of the mixed raw materials through the image analysis on the corresponding images It can also be acquired automatically. In addition, for example, the operation data collecting unit 110 photographs the sintered ores emitted from the sintering machine 30 to the light distribution unit through at least one camera in real time, analyzes the image data, May be automatically obtained. In this case, the camera for photographing the sintered ores can be arranged to photograph the light distribution portion for conveying the sintered ores discharged from the sintering machine 30 to the blast furnace.

The operation data collecting unit 110 may receive the operation data from the operator through the user input unit 120.

In addition, the operation data collecting unit 110 may receive operation data from an external facility.

The operation data collecting unit 110 may store the operation data collected in the operation data database 130 in a time series.

The learning unit 140 includes a neural network algorithm based prediction model for learning operational data collected through the operation data collecting unit 110 for training data for learning and for predicting operation guidance related to the sintering operation, Can be generated.

The neural network algorithm used for generating the predictive model in the learning unit 140 is composed of a convolutional neural network (CNN). CNN is a neural network consisting of one or several convolutional layers, a pooling layer, and fully connected layers. In CNN, features are extracted from time series data through a convolutional layer and a pooling layer, and prediction is performed through classification in a fully connected layer.

In the embodiment of the present invention, the learning unit 140 may set the number of features per variable to four and the number of time series data to eight (8 equal) in order to generate a predictive model. In addition, the fully connected layer is composed of two layers, and the number of neurons in each layer is set to 100 and 50, so that learning can be performed. The predictive model thus generated can predict an operation value including at least one of the combination ratio in the blended raw material, the post-finishing ratio, the charging density of the blended raw material, and the moving speed of the sintering machine 30 from the inputted time series operational data have.

When the prediction model is generated, the learning unit 140 may perform the consistency and reliability verification process for the prediction model to verify the validity of the operation value provided by the prediction model.

The learning unit 140 can perform consistency and reliability verification for the model to be checked by comparing the manipulated values in the actual sintering operation with the manipulated values obtained using the predictive model.

For example, as shown in FIG. 3, the target value of the predictive model is set as the sintered normal-temperature intensity, and the intensity of the sintered normal-temperature strength in the actual sintering operation is compared with the intensity of the sintered ordinary- It is possible to perform verification of consistency and reliability. For example, in the case where the room temperature strength of the actual sintering light is higher than the operating standard value and the coupling ratio or the charging density should be lowered, the learning unit 140 also predicts the room temperature strength of the sintered ores at a high level to lower the bonding cost and the charging density It is possible to verify the consistency and reliability of the prediction model by confirming whether the prediction result is presented.

The learning unit 140 may automatically control the sintering operation using a prediction model and monitor the operation efficiency (chartering amount) and the operation stabilization state (deviation of the management indicator) to verify the consistency and reliability of the prediction model .

The prediction model generated by the learning unit 140 is stored in the prediction model database 150 and used for predicting the operation value related to the sintering operation in the prediction unit 160. [

The predicting unit 160 can estimate operation values related to the sintering operation from the operation data, which are time series data, using a neural network algorithm based prediction model. That is, the predicting unit 160 inputs the operation data collected through the operation data collecting unit 110 as the time series input data of the prediction model based on the neural network algorithm, and outputs the output value of the prediction model to the operation values .

The sintering operation control unit 170 can automatically control the sintering facility based on the operation values output by the predicting unit 160. [ For example, the sintering operation control unit 170 controls the amount of the binder material supplied from the storage bin (not shown) to the mixer for raw material mixing (not shown) based on the combined raw material ratio in the mixing raw material among the predicted operation values, Can be controlled. Further, for example, the sintering operation control unit 170 controls the mixing material feed amount of the mixing material storage hopper 10 on the basis of the post-lamination operation of the predicted operating values, It is possible to adjust the amount. For example, the sintering operation control unit 170 controls the pressing member (not shown) located on the sintering machine 30 and the like to set the charging density of the sintering compounding raw material Can be adjusted. Further, for example, the sintering operation control unit 170 can control the moving speed of the sintering machine 30 by controlling the conveyor 40 on the basis of the moving speed of the predicted operating values.

The functions of the operation data collecting unit 110, the learning unit 140, the predicting unit 160 and the sintering operation controlling unit 170 may be performed by one or more central processing units unit, CPU) or other chipset, microprocessor, or the like.

4 is a schematic view illustrating a sintering operation control method of a blast furnace according to an embodiment of the present invention.

Referring to FIG. 4, a sintering operation control apparatus 100 according to an embodiment of the present invention generates a predictive model through neural network algorithm, for example, CNN-based learning (S100).

In operation S100, the sintering operation control apparatus 100 may collect and accumulate operation data of the sintering facility for a predetermined period of time, and generate a prediction model by using accumulated operation data of the time series as learning data of the neural network algorithm .

As the prediction model is generated, the sintering operation control apparatus 100 continuously collects a plurality of operation data related to the operation state of the sintering facility for operation values related to the sinter operation (S110). Then, operation data related to the sintering operation is obtained (S120) by using the continuously collected operation data of the time series as the time series input data of the prediction model based on the neural network algorithm.

Further, when the operation value related to the sintering operation is obtained, the sintering operation control apparatus 100 automatically controls the sintering facility based on the operation value (S130).

In step S130, the sintering operation control device 100 controls the amount of binder supplied from the storage bin (not shown) to the mixer for raw material mixing (not shown) based on the combined ratio of the raw materials in the mixing operation, It is possible to control the binding ratio of the raw material mixture. The sintering operation control apparatus 100 may also control the amount of the compounding material accumulated on the bed of each sintering machine 30 by controlling the compounding material storage hopper 10 on the basis of the post-deposition of the predicted operation values have. The sintering operation control apparatus 100 may also control the charging density of the blending raw material for sintering by controlling a pressing member (not shown) located above the sintering machine 30 based on the charging density of the predicted operating values . Also, the sintering operation control device 100 may control the moving speed of the sintering machine 30 by controlling the conveyor 40 based on the moving speed of the predicted operating values.

According to the above-described embodiment, the sintering operation control device 100 predicts the operation values corresponding to the current sintering operation state in real time using the neural network-based prediction model, and automatically controls the sintering facility based on the predicted operation values. Therefore, it is possible to control the sintering equipment in real time according to the current operating conditions, thereby improving the efficiency of the sintering operation and stably maintaining the quality of the sintered ores.

The sintering operation control method according to the embodiment of the present invention can be executed through software. When executed in software, the constituent means of the present invention are code segments that perform the necessary tasks. The program or code segments may be stored on a computer readable recording medium.

A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording device include ROM, RAM, CD-ROM, DVD-ROM, DVD-RAM, magnetic tape, floppy disk, hard disk and optical data storage device. Also, the computer-readable recording medium may be distributed over a network-connected computer device so that computer-readable code can be stored and executed in a distributed manner.

It is to be understood that both the foregoing general description and the following detailed description of the present invention are illustrative and explanatory only and are intended to be illustrative of the invention and are not to be construed as limiting the scope of the invention as defined by the appended claims. It is not. Therefore, those skilled in the art can readily select and substitute it. Those skilled in the art will also appreciate that some of the components described herein can be omitted without degrading performance or adding components to improve performance. In addition, those skilled in the art may change the order of the method steps described herein depending on the process environment or equipment. Therefore, the scope of the present invention should be determined by the appended claims and equivalents thereof, not by the embodiments described.

(Explanation of Symbols)

10: Mixing raw material storage hopper

20: By ignition

30: Sintering machine

40: Conveyor

50: Wind Box

61: Duct

62: Dust collector

63: Duct blower

70: Hood

80: fuel storage portion

100: Sintering operation control device

110: Operation data collecting unit

120: user input section

130: Operation data database

140:

150: Predictive model database

160: prediction unit

170: Sintering operation control section

Claims

1. A sintering operation control device comprising:

A database storing a prediction model for predicting at least one manipulation value associated with the sintering operation through a neural network algorithm,

Acquiring a particle size distribution of the sintered ores from an image of the sintered ores emitted from the sintering machine using at least one camera and collecting operational data including the particle size distribution in relation to the operating state of the sintering operation part,

A prediction unit that uses the operation data as input data of the prediction model to obtain the at least one operation value;

And a control unit for automatically controlling the sintering equipment in accordance with the at least one operation value,

Wherein the operation data includes particle size distribution data of sintered ores.
The method according to claim 1,

Wherein the at least one operating value includes at least one of a sintering operation including a mixing ratio in a blending raw material charged into a sintering machine, a blending raw material height on a bed of the sintering machine, a blending raw material loading density of the sintering machine, controller.
3. The method of claim 2,

Wherein the control unit controls the amount of the binder to be supplied from the storage bin to the mixer for raw material mixture based on the combination ratio in the blended material among the at least one operation value.
3. The method of claim 2,

Wherein the control unit controls the mixing material feed amount of the mixing material storage hopper based on the mixing material height on the bed of the sintering machine among the at least one operation value.
3. The method of claim 2,

Wherein the control unit controls the pressing member on the sintering unit based on the mixing material loading density of the sintering machine among the at least one operating value.
3. The method of claim 2,

Wherein the control unit controls a conveyor that moves the sintering machine based on a moving speed of the sintering machine among the at least one manipulated value.
The method according to claim 1,

Wherein the neural network algorithm comprises a convolution neural network.
The method according to claim 1,

Further comprising a learning unit configured to learn the operation data collected through the operation data collecting unit with data for training for a predetermined period to construct the prediction model.
The method according to claim 1,

The operating data includes at least one of a raw material mixture ratio, a blending raw material component, a coupling ratio in a blending raw material, a binder particle size, an ignition temperature and a gas pressure in an ignition, a moisture content of the blending raw material, a dust generating amount, The flow rate of the sintering machine, the flow speed of the sintering machine, the charging density, the air permeability, the exhaust gas temperature, the oxygen concentration, the negative pressure and flow rate of the hood, the sintered light component, the sintered light particle size, the sintered light intensity, Wherein the sintering operation control device further includes at least one sintering operation control device.
In the sintering operation control method,

Obtaining a particle size distribution of the sintered ores from an image of the sintered ores emitted from the sintering machine using at least one camera,

Collecting operational data including the particle size distribution in relation to the operational state of the sintering operation,

Using the operational data as input data of a predictive model predicting at least one manipulated value associated with a sintering operation through a neural network algorithm to obtain the at least one manipulated value,

And automatically controlling the sintering equipment in accordance with the at least one operating value.
11. The method of claim 10,

Wherein the at least one operating value includes at least one of a binding material ratio in the blended material charged into the sintering machine, a blending material height on the bed of the sintering machine, a mixing material loading density of the sintering machine, and a balancing speed of the sintering machine / RTI >
12. The method of claim 11,

Wherein the controlling comprises:

And controlling an amount of the binder to be supplied from the storage bin to the mixer for raw material, based on the binder ratio in the blend material among the at least one operation value.
12. The method of claim 11,

Wherein the controlling comprises:

And controlling the feed amount of the blending raw material storage hopper based on the blending raw material height on the bed of the sintering machine among the at least one operating value.
12. The method of claim 11,

Wherein the controlling comprises:

Controlling the pressing member on the sintering machine based on the mixing material loading density of the sintering machine among the at least one operating value.
12. The method of claim 11,

Wherein the controlling comprises:

And controlling a conveyor for moving the sintering machine based on the moving speed of the sintering machine among the at least one operation value.
11. The method of claim 10,

Wherein the neural network algorithm comprises a convolution neural network.
11. The method of claim 10,

Further comprising the step of: learning the operation data with learning data for a predetermined period to construct the prediction model.