WO2019107980A1

WO2019107980A1 - Quality prediction device and method

Info

Publication number: WO2019107980A1
Application number: PCT/KR2018/014981
Authority: WO
Inventors: 유종우; 조병국; 정은호; 박종인; 김병일
Original assignee: 주식회사 포스코
Priority date: 2017-11-30
Filing date: 2018-11-29
Publication date: 2019-06-06
Also published as: KR20190063836A; KR102074359B1

Abstract

The present invention presents a quality prediction device and a quality prediction method employing same, the quality prediction device comprising: a storage unit for turning state information and quality information of a treatment subject into data and accumulating same; a measurement unit for acquiring current state information of the treatment subject, in a treatment subject treatment path; and a quality prediction unit for extracting, based on the data accumulated in the storage unit, a data set corresponding to the current state information, and then predicting the quality of the treatment subject by using same. The present invention presents a quality prediction device and method which can turn treatment performance of a treatment subject into data and accumulate same, and accurately predict the quality of the treatment subject on the basis of the accumulated data.

Description

Apparatus and method for predicting quality

The present invention relates to a quality prediction apparatus and method, and more particularly, to a quality prediction apparatus and method capable of accurately estimating the quality of an object to be processed based on accumulated data by accumulating and accumulating processing results of the object to be processed will be.

Sintered ores are blast furnace materials made of iron ore, limestone, coke, and anthracite as raw materials and manufactured to a size suitable for blast furnace use. The sintered ores are produced through a process of preparing a raw material mixture and a process of sintering the raw material mixture. Among them, the sintering process is usually carried out in a dewatering sintering machine.

The sintering machine is a machine in which the bogie is moved in the order of charging, ignition, sintering and cooling sections, the raw material layer is formed by charging the raw material mixture into the bogie, the combustion bands are formed on the top of the raw material layer, Is moved from the upper part of the raw material layer to the lower part, and the raw material layer is sintered and cooled to produce a sintered light in the form of a cake. Thereafter, the sintering machine is made from sintered light by distributing the sintered cake from a truck, crushing and cooling it.

The sintered ores are classified into a particle size of several to several tens of millimeters, which is a size suitable for blast furnace use, and are transported to the blast furnace. The sintered ores are used as a raw material in a sintering process (hereinafter referred to as a "blast furnace process") carried out in a blast furnace. Therefore, the quality of the sintered ores is an essential condition for the stabilization of the sintering process.

It takes about 30 minutes for the bogie to pass through each section of the sintering machine, and the quality of the sintered ores is determined within this time. However, at this time, the quality of the sintered ore can not be directly measured. For example, when the sintered cake is shredded in a truck, and the sintered cake is crushed using a hot crusher, the crushed sintered cake is charged into the cooler and allowed to stay for about 90 minutes and cooled. Thereafter, when selecting the size of the sintered ores cooled in the multistage screen, the quality of the sintered ores and the strength of the sintered ores are measured by using a sample after collecting the sintered ores for quality confirmation. As such, the sinter ore can be sampled and measured for quality only after about two hours have passed since the quality was determined.

The method of operating the sintering machine is as follows. First, the operation of the sintering machine is started with the established operating conditions, and then the quality of the sintered ores is monitored, and the operation conditions are modified with the result, and the operation of the sintering machine is continued. However, as described above, there is a considerable time gap between the point of time at which the sintered ores are shined and the point at which the quality of the sintered ores is measured.

Therefore, in order to overcome the temporal blank, the operator estimates the quality of the sintered ore with each criterion, modifies the operating conditions according to the result, and reflects the modifications to the operation of the sintering machine first. That is, since the operator directly adjusts the operating conditions according to their respective standards until the measurement of the sinter ore quality, continuity of the process when the operator alternates can not be guaranteed. Also, due to the limited manpower situation, if the operator fails to continuously observe the sintering machine and misses an appropriate response time, it may cause a quality deviation.

Techniques as a background of the present invention are listed in the following patent documents.

(Prior art document)

(Patent Literature)

(Patent Document 1) JP2013-151715 A

(Patent Document 2) KR10-1442983 B1

The present invention provides a quality predicting apparatus and method capable of accurately accumulating and accumulating processing results of an object to be processed and accurately estimating the quality of the object to be processed based on the accumulated data.

The present invention provides an apparatus and method for predicting the quality of an object to be processed using an artificial intelligence-based in-depth neural network.

A quality predicting apparatus according to an embodiment of the present invention includes: a storage unit for accumulating and accumulating data of a quality of an object to be processed associated with state information and state information of an object to be processed; A measurement unit for acquiring current state information of the object to be processed in the object processing path; And a quality predicting unit for extracting a data set corresponding to the current state information based on the data stored in the storage unit and predicting the quality of the object using the data set.

Wherein the measuring unit comprises: a first measuring unit for obtaining an exhaust gas temperature distribution of the object to be processed along the extending direction of the path; And a second measuring unit for obtaining a temperature distribution in the object to be processed in a direction across the path at an end point of the path.

The first measuring unit and the second measuring unit may obtain the state information by patterning in the form of thermal image information.

The storage unit may store cause information about the quality information together.

The storage unit may store the status information, the corresponding quality information, and the cause information for the quality information, and store the resultant information in a data set.

And a control unit connected to the quality predicting unit, wherein the control unit can control at least a part of the object treatment facility using the cause information matching the predicted quality.

The quality predicting unit can extract a data set using the artificial intelligence based depth network and predict the quality of the object to be processed.

The artificial intelligence may include deep running, and the in-depth neural network may include a convolutional neural network.

A quality predicting method according to an embodiment of the present invention includes a process of accumulating and accumulating data on the quality of an article to be processed associated with state information and state information of a subject to be processed; Processing the object to be processed in the processing path; Acquiring current state information of the object to be processed in the processing path; Extracting a data set corresponding to the current state information from the accumulated data and using the data set to predict the quality of the object to be processed; And outputting the current state information and the predicted quality according to the current state information.

Wherein the step of acquiring the state information comprises: patterning the exhaust gas temperature distribution of the object to be processed in the form of thermal image information in the direction of the extension of the path; And patterning the temperature distribution in the object to be processed in the form of thermal image information in the direction crossing the path at the end point of the path.

It is possible to store the matching result as a data set after matching the cause information of the quality information in addition to the quality information of the object to be processed associated with the status information and the status information.

The step of extracting the data set may include the steps of: comparing the current state information with state information in the accumulated data using a neural network based on artificial intelligence; Selecting state information in the accumulated data corresponding to the current state information, and extracting a data set including the selected state information; And predicting quality information of the object to be processed included in the extracted data set with the current quality of the object to be processed.

If there is no state information matching the current state information in the accumulated data, a data set including state information most similar to the current state information can be extracted in the accumulated data.

Outputting the predicted quality, if the predicted quality is output as fluctuation and abnormality, together with the cause information; And controlling at least a part of the object treatment facility using the output cause information.

Acquiring actual quality information of the object to be processed associated with the current state information after the process of predicting the quality of the object to be processed; And adding the current state information and the actual quality information to the accumulated data.

The object to be treated includes sintered ores, and the treatment path may include a sintering section through which a sludge passage of the sintering machine can pass.

According to the embodiment of the present invention, the processing performance of the object to be processed can be recorded and accumulated, and the quality of the object to be processed can be accurately predicted based on the accumulated data using the artificial intelligence-based in-depth neural network.

For example, when the present invention is applied to a sintering process (hereinafter referred to as "sintering process") performed in a sintering machine, various processing results of the sintering process are accumulated and stored, and a deep learning based convolutional neural network ) Is used to extract a data set including the state information similar or identical to the current sintered-state information in the accumulated data. The quality information of the sintered ores contained in the extracted data set can be accurately predicted by the sintering quality of the current sintering process.

In other words, by objectively predicting the quality of the sintered ore without delay in the process and providing accurate guides for correcting the operating conditions of the sintering facility based on the result, it is possible to prevent fluctuations in the quality of the sintered ore and ensure uniform sintering quality .

In addition, the sintering machine can be rationally operated by an accurate guide, and it is possible to reduce the cost of the joint, reduce the amount of the by-product and the carbon dioxide gas, and reduce the cost.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic view of a quality predicting device and an object to be treated according to an embodiment of the present invention. Fig.

2 is a diagram showing an example of data accumulated according to an embodiment of the present invention.

FIG. 3 is a diagram illustrating an example of a process of extracting a data set using a neural network according to an embodiment of the present invention. Referring to FIG.

4 is a flowchart of a quality estimation method according to an embodiment of the present invention.

5 is a graph for explaining the accuracy of a quality prediction result according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to the embodiments described below, but may be embodied in various forms. It is to be understood that both the foregoing general description and the following detailed description of the present invention are exemplary and explanatory and are intended to provide further explanation of the invention as claimed. BRIEF DESCRIPTION OF THE DRAWINGS The drawings may be exaggerated for purposes of describing embodiments of the present invention, wherein like reference numerals refer to like elements throughout.

An apparatus and method for predicting quality according to an embodiment of the present invention can accumulate processing data of an object to be processed and can provide a technical feature capable of accurately predicting the quality of the object to be processed based on the accumulated data.

The apparatus and method for predicting quality according to an embodiment of the present invention are applied to a sintering process of a steel mill and can be applied to various other treatments. Hereinafter, embodiments of the present invention will be described with reference to a sintering process.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic view showing a quality predicting device and an object to be treated in accordance with an embodiment of the present invention. Fig. 1, the object to be treated according to an embodiment of the present invention includes a hopper 10 (hereinafter referred to as a " processing path " (20) which is spaced apart from the hopper (10) in the direction in which the object to be treated (hereinafter referred to as the "processing step") proceeds, Batches (20) for receiving the blend material from the hopper (10) and loading it therein, windboxes (40) installed in the lower part of the processing path and communicating with the inside of the bogies (30) A dust collector 82 provided on the other side of the duct 81 and a duct 81 connected to the dust collector 82 and a duct 81 connected to the dust collector 82, And a blower 83 installed therein.

The compounding raw material may be a raw material for producing the object to be treated. For example, the blending raw material may be a raw material for producing the sintered ores. In this case, the blending raw materials may include minute iron ore, limestone, minute coke and anthracite. The material to be treated may be an sintered ore, and the treatment process may be a sintering process. Of course, this need not be particularly limited.

The hopper 10 is installed on a loading zone to be described later. The hopper 10 is provided with a drum feeder and a chute at a lower portion thereof. The drum feeder and the chute load the mixing material in the hopper 10 vertically segregated in the inside of the carriage 30. [ A material layer to be described later is formed in a charging zone.

The blended raw materials charged into the carts 30 are referred to as raw material layers. The raw material layer may be referred to as a sintered bed. The raw material layer is heat-treated, for example, sintered while moving the treatment path in the direction in which the treatment process proceeds, and is distributed in the form of a sintered cake in the light-shielding portion, which is the end point of the treatment path.

The ignition furnace 20 is installed on the ignition section to be described later. The ignition furnace 20 can be supplied with fuel. The fuel is burned in the ignition furnace 20 to generate a flame and the flame is injected into the inside of the bogies 30. [ The flame is ignited on the surface of the raw material layer to form a combustion zone. That is, a combustion zone is formed in the ignition zone.

In the sintering section to be described later, the combustion zone can move from the surface of the raw material layer to the lower layer via the upper layer. During the combustion zone movement, the sintering reaction, which is a reaction in which limestone and mined iron ores form a low melting point compound, proceeds near the combustion zone. By this reaction, the raw material layer can be sintered into sintered ores. That is, the raw material layer is sintered in the sintering section.

The bogies (30) are opened upward, and openings are formed at the bottom. Through the openings, the inside of the bogies (30) communicates with the wind boxes (40). The inside of the carriages 30 is sucked downward by the wind boxes 40. [ Bogies 30 are connected to each other endlessly. A conveyor may be installed to support the carts 30. For example, a conveyor is installed in the direction in which the treatment process proceeds, and the conveyors 30 are installed. The upper side of the conveyor forms a processing path, and the lower side of the conveyor forms a conveyance path. The carts 30 can travel on the treatment path in the direction in which the treatment process proceeds. Then, the carts 30 pass through the light distribution portion, distribute the object to be processed, such as sintered light, in the form of a cake, and enter the return path. Then, the carts 30 can travel on the return path in the direction opposite to the direction in which the treatment process proceeds, and return to the starting point of the treatment path.

The processing path A includes a plurality of sections. The plurality of sections include a charging section B, an ignition section C, and a sintering section D. At this time, it may be the order of the charging zone B, the ignition zone C, and the sintering zone D in the direction in which the treatment process proceeds. The raw material layer can be sintered while passing through the sintering section (D). Meanwhile, the sintering section D includes a first section D1, a second section D2 and a third section D3. The criteria for dividing these sections may vary. For example, it may be based on the extension length of the sintering section D, on the basis of the point where the exhaust gas temperature is maximum, or on the point where the combustion of anthracite coal, which is solid fuel in the raw material layer, is terminated. In addition, the criteria may vary, and it is not necessary to limit it in the examples in particular.

For example, the sintering section D is defined as a first section D1 from the starting point of the sintering section D to the first half of the first half of the sintering section D and the third section from the end point of the sintering section D to the third half (D3), and an intermediate portion between the first section D1 and the third section D3 may be defined as the second section D2.

Windboxes (40) provide suction force downward into the carriages (30). At this time, the wind boxes 40 can form a negative pressure therein by the blower 83, and the inside of the bogies 30 can be sucked downward by using the blowers. The exhaust gases are collected into the windboxes 40. The exhaust gas may be recovered to the duct 81.

When the object to be treated (also referred to as "sintering machine") is operated, firstly, the operation of the object to be treated is started with the established operation conditions. For example, prior to commencement of the operation of the object treatment facility, a physical model of the sintering process is used based on the information related to the material mixture to establish operating conditions for achieving the desired quality of the sinter ores, The operation of the facility is started. At this time, the information related to the blended raw materials may be various, such as raw material composition, binding cost, and anthracite ratio. In addition, the operating conditions may vary, such as bed speed, bedding, and loading density.

Thereafter, the quality of the sintered ores is monitored, the operating conditions are modified with the result, and the operation of the sintering machine is continued.

On the other hand, the quality of the sintered ores is determined in the sintering section. Therefore, if the quality of the sintered ores is precisely predicted at the time when the sintered ores are lighted, for example, with the state information of the sintered ores, the operating condition of the sintering machine can be modified immediately so that the quality of the sintered ores can be maintained at a desired level.

To this end, the object to be treated according to the embodiment of the present invention may include a quality predicting device. At this time, the quality prediction apparatus monitors the state information representing the quality of the sintered ore, and can accurately predict the quality of the sintered ores.

And the quality predicting device can continuously operate the object treatment facility with the modified operation condition while correcting the operation condition by reflecting the quality prediction result continuously or periodically to the operation condition.

Referring to FIG. 1, a quality prediction apparatus according to an embodiment of the present invention includes a storage unit 600 for storing and accumulating quality information of an object to be processed associated with status information of the object to be processed and status information, A quality estimating unit 700 for extracting a data set corresponding to the current state information based on the data stored in the storage unit and estimating the quality of the object to be processed using the data set, . The quality predicting apparatus may further include a controller (not shown) connected to the quality predicting unit 700. The control unit can control at least a part of the object treatment facility using the cause information matching the predicted quality. That is, the quality prediction apparatus can function as a control apparatus for operating the object to be treated. Therefore, the quality prediction apparatus according to the embodiment of the present invention may be referred to as a control apparatus of the object to be treated.

The storage unit 600 may track the quality information of the object to be processed associated with the state information of the object to be processed and the state information, and may accumulate the quality information together with the state information. Also, the storage unit 600 may track the cause information about the quality information of the object to be processed and store them together. At this time, the storage unit 600 may store the status information, the corresponding quality information, and the cause information of the quality information, and accumulate them as a data set.

For example, if the state information of a predetermined number of times is stored in the storage unit 600 while collecting the information for each time of the sintering process which is repeated several tens to several hundred times or more and continuously accumulating the data set of the total number of times, The cause information of the same rotation as the information is matched with this state information and is stored together.

2, the state information (third information) of the object to be processed includes, for example, the exhaust gas temperature distribution obtained in the third section D3 of the sintering section and the temperature distribution in the object to be processed obtained in the light-distribution section. Herein, the exhaust gas temperature distribution means the exhaust gas temperature distribution obtained inside the wind boxes of the third section of the wind boxes 40. The temperature distribution in the material to be treated means the distribution of the glowing layer at the cross section of the sintered cake. The state information of the object to be processed is information that can be obtained from the sintering machine and is stored as "firing pattern (S1, S2 ...)" in the accumulated data. The storage unit 600 patterns the plastic patterns in the form of thermal image information and stores them. On the other hand, in the drawing, the x-axis is the traveling direction of the treatment process, the y-axis is the width direction of the treatment path, and the z-axis is the height direction of the raw material layer. 2, the right portion of the thermal image is on the light pipe side, and the left side portion is on the upstream side of the light pipe portion, for example, the ignition path 20 side. The temperature on the upstream side of the light-transmitting portion is relatively low and the temperature on the light-emitting portion side is relatively high. Further, the temperature gradually increases from the upstream side of the light-directing portion to the light-directing portion side. Such a temperature change state or pattern can be obtained in the form of thermal image information of the exhaust gas temperature distribution.

The quality information (fourth information) of the article to be processed associated with the state information of the article to be treated may include the intensity, particle size, productivity of the sintered ores and the current state (exit ratio and reduction ratio) in the blast furnace process using the sintered ores. The strength, particle size and productivity among the quality information of the object to be treated can be obtained by sampling the sintered ores after about two hours after the light distribution of the sintered ores, and the status in the blast furnace process can be obtained from a blast furnace process which is a subsequent process. The acquired information is stored as "output condition (O1, O2 ...)" in the accumulated data.

The cause information of the quality information of the object to be treated is information obtained by tracking the cause of the quality of the object to be treated, and there are roughly two kinds of information. One of them is information related to the raw material of the blend, and the other is information related to the sintering machine. The mixing material information (referred to as 'first information') includes, for example, raw material components / particle size, binding ratio / particle size, water / quicklime / anthracite ratio and the like. ) &Quot;. Further, the sintering machine information (referred to as 'second information') includes, for example, a bogie speed, a layer thickness, a loading density, an air flow rate and the like and is stored in the accumulated data as the "operating condition P1, P2.

The above-mentioned information collectively refers to the processing performance of the object to be processed. The accumulated data is, for example, a set of huge experimental data having input conditions, operating conditions, firing patterns, and output conditions, and the quality prediction unit 700 uses the data as data for predicting the quality of the material to be processed.

On the other hand, in the embodiment, it is judged that the state information of the object to be processed can represent the quality of the object to be processed, and the quality of the object to be processed is predicted by utilizing the state information. The reason for this is as follows.

For example, it is best if the quality information is predicted only by the input condition, but since the output condition is influenced by the operation condition, it is impossible to predict the quality information only by the input condition. If the firing result of the sintering process varies depending on the input conditions and the operating conditions, such fluctuations are immediately expressed in the exhaust gas temperature distribution and in the distribution of the heat radiation layer. That is, both the input conditions and the operating conditions are reflected in the firing pattern.

Therefore, if state information of the object to be processed, which is a plastic pattern, among the processing results in the accumulated data is used as an index for quality prediction, the quality of the object to be processed can be predicted quickly before sampling of the object to be processed.

At this time, the quality predicting unit 700 uses the artificial intelligence-based depth network to improve the accuracy of prediction. Therefore, objective predictions can be made by the learned neural network based on the accumulated data, rather than being subjectively predicted by the operator with the state information of the object to be processed.

The measuring unit can obtain the current state information of the object to be processed in the processing path. The measuring section includes a first measuring section 510 for obtaining an exhaust gas temperature distribution of the object to be processed along the extending direction of the processing path and a second measuring section 510 for determining the temperature distribution within the object to be processed And a second measurement unit 520 for acquiring the second measurement value.

The first measuring unit 510 may be a thermometer, for example, and may be disposed in the third section D3. The first measuring unit 510 may be disposed in, for example, the wind boxes 40, which is a position at which the temperature of the exhaust gas can be measured, but is not limited thereto. The first measuring unit 510 is disposed at five to fifteen positions spaced apart from each other in the proceeding direction of the processing process and is disposed at five to ten positions spaced apart in the width direction of the processing path, As shown in FIG. Of course, their spacing and number may vary. The first measurement unit 510 can be used to obtain the temperature distribution of the exhaust gas passing through the combustion zone in the horizontal direction. The temperature of the flue gas varies depending on the sintering condition of the sintered ores, and the horizontal sintering pattern information of the sintered ores can be obtained through the temperature distribution of the flue gas.

The second measuring unit 520 may be a thermal imager or a CCD image sensor. The second measuring unit 520 is disposed outside the light-directing unit and is provided so as to be able to shoot an end face of the sintered cake to be light-distribution. By using the second measuring unit 520, the temperature distribution in the object to be treated can be obtained in detail in the cross section of the sintered cake photographed by the light distribution unit. This is also referred to as information on the vertical sintering state of the sintered ores. The meaning of the glow layer is as follows. For example, when the sintering process is performed while controlling the thickness of the heat-generating layer to a certain range, the quality of the sintered ores can be obtained. That is, the thickness control of the glowing layer is related to the quality of the sintered ores.

The first measuring unit 510 and the second measuring unit 520 are connected to the storage unit 600. The first measuring unit 510 and the second measuring unit 520 pattern the state information in the form of thermal image information and acquire the information, Output.

The data stored in the storage unit 600 is data for predicting the current state information of the object to be processed. For example, if the current state information of the object to be processed is measured by the measuring unit and the state information similar or identical to the current state information is found in the accumulated data, the quality information matching the found state information may be output to predict the current state information .

That is, it is possible to find a similar firing pattern in the accumulated data, which is the past performance by way of the currently obtainable firing pattern, and predict the performance expected in the current process as various pieces of performance information matching the firing pattern. The quality predicting unit 700 performs this.

The quality predicting unit 700 may be connected to the storage unit 600. The quality predicting unit 700 can extract the data set corresponding to the current state information based on the data stored in the storage unit 600 and predict the quality of the object to be processed using the data set.

The quality predicting unit 700 can extract the data set using the artificial intelligence-based in-depth neural network, and predict the quality of the object to be processed. At this time, the artificial intelligence includes deep running, and the deep neural network can include convolutional neural networks. In other words, high accuracy of predictions can be realized by utilizing the artificial intelligence algorithms that are currently developed, and the quality information can be predicted preliminarily and objectively using a formalized plastic pattern. The quality predicting unit 700 will be described in detail below when describing the quality predicting method.

A control unit (not shown) is connected to the quality predicting unit. The control unit outputs the predicted quality in a variety of ways that the operator can perceive, such as a text or graphic on the screen, or a voice. In addition, the control unit controls at least a part of the object treatment facility using the cause information matching the predicted quality. The control unit can control the hopper 10, the ignition path 20, the carts 30, and the blower 83. [ Of course, the control section can control the overall operation of the object to be treated. The controller can be used to notify the operator of the disturbance factor of the quality fluctuation, and to control the disturbance factor by controlling the processing facility. The control by the control unit is not particularly limited. For example, there may be a variety of control methods such as controlling the dispensing rate of the raw material mixture in the hopper 10, controlling the speed of the carts 30, or adjusting the amount of fuel supplied to the ignition passages 20. Meanwhile, the control unit may control the operation of the measurement unit, the storage unit 600, and the quality prediction unit 700.

FIG. 3 is a diagram illustrating an example of a process of extracting a data set using a neural network according to an embodiment of the present invention, and FIG. 4 is a flowchart of a quality estimation method according to an embodiment of the present invention. 5 is a graph for explaining the accuracy of the quality prediction result according to the embodiment of the present invention.

Hereinafter, a quality prediction method according to an embodiment of the present invention will be described with reference to FIGS. 1 to 5. FIG.

The quality prediction method according to an embodiment of the present invention includes a step S100 of accumulating and accumulating the quality information of the object to be processed related to the state information and the state information of the object to be processed, (S400) of extracting a data set corresponding to the current state information from the accumulated data and estimating the quality of the object to be processed using the extracted data set, And outputting the current state information and the predicted quality accordingly (S500). At this time, as described above, the object to be treated includes sintered ores, and the treatment path includes a sintering section through which the sludge of the sintering machine can pass.

S100: The state information of the object to be processed and the quality information of the object to be processed associated with the state information are converted into data and stored in the storage unit 600. At this time, after matching the cause information of the quality information in addition to the quality information of the object to be processed associated with the status information and the status information, the matching result is accumulated as a data set. The contents thereof have been described in detail when explaining the quality predicting apparatus, and thus, are omitted in order to avoid duplication of explanation. This process can be carried out continuously by repeating the sintering process. If a certain amount of data is stored first, quality can be predicted in the following process.

S200: Process the object to be processed in the processing path. That is, the sintering process is performed. The compounding materials are charged into the carts 30, and the carts 30 are sintered while moving to the processing path to produce sintered ores.

S300: obtains the current state information of the object to be processed in the processing path. The temperature distribution of the exhaust gas of the article to be processed is obtained by using, for example, a measuring section, and the temperature distribution in the article to be processed, that is, the distribution of the sintered cake cross- At this time, the exhaust gas temperature distribution of the object to be processed is patterned and obtained in the form of thermal image information in the direction of extension of the path, and the temperature distribution in the object to be processed in the direction crossing the path at the end point of the path, And obtained by patterning.

S400: The data set corresponding to the current state information is extracted from the accumulated data, and the quality of the object to be processed is predicted using the data set. Referring to FIG. 3, the present state information is compared with state information in the accumulated data by using the artificial intelligence-based in-depth neural network. At this time, the pattern of the inputted current state information is recognized, and a pattern of state information in the data is recognized, and then a matching pattern is found. This process can be performed using a deep learning based convolutional neural network, thus increasing its accuracy. Upon completion of the comparison process, state information in the accumulated data matching the input current state information is selected, and a data set (also referred to as a "data set") containing the selected state information is extracted. As shown in the figure, various performances are matched in the data set, and the quality information of the object to be processed included in the extracted data set is predicted with the current quality of the object to be processed.

On the other hand, if there is no state information matching the current state information input in the accumulated data, it is possible to extract a data set including the state information most similar to the inputted current state information in the accumulated data. In addition to this, the method of extracting the data set by comparing the status information may be various. Although only the flue gas temperature distribution is shown in the figure, the glow layer distribution is also used in the comparison process.

S500: Thereafter, the control unit is used to output the input current state information and the predicted quality accordingly. The output method may be various, and is not particularly limited. For example, when outputting quality, the quality value is out of the range based on a predetermined range. If the quality value is within a threshold value of the range, the variation is output. If the quality value is out of the range and the value exceeds the threshold value, do. Here, the threshold value means quality fluctuation to such an extent as to influence the present state of the blast furnace process. The threshold can be set to any value in conjunction with the blast furnace performance accumulated through previous processes.

On the other hand, if the predicted quality is output as fluctuation and abnormality after the process of outputting the predicted quality, the cause information about the quality is outputted together. Then, at least a part of the object treatment facility is controlled by the control unit using the outputted cause information. By this process, the quality of the object to be processed can be quickly returned to a desired quality.

After the process of predicting the quality of the object to be processed, actual quality information of the object to be processed associated with the current state information is obtained. For example, the produced sintered ores are sampled to obtain actual quality information. Then, the current state information and the actual quality information are converted into data and added to the accumulated data. Thus, the accumulated data can be continuously reinforced as it becomes richer.

Referring to FIG. 5, a sieving process is performed and an algorithm of a deep learning-based convolution neural network is applied to predict the quality of an object to be processed, actual quality is tracked, And the error is shown.

The horizontal axis of the graph in Fig. 5 represents the error of the drop strength, and the vertical axis represents the number of cases. When the drop strength error is within ± 0.4, the accuracy is about 61% and within ± 1.0, it is 90%. Accuracy can be improved if more data is accumulated and learning algorithms are added. As an artificial intelligence algorithm applied to the embodiment, it can be seen that CNN which is strong in pattern recognition is suitable from general machine learning algorithm according to the quality and number of quality information.

The above-described embodiments of the present invention are for the explanation of the present invention and are not intended to limit the present invention. It should be noted that the configurations and the methods disclosed in the above embodiments of the present invention may be modified into various forms by combining or intersecting with each other, and such modifications may be considered within the scope of the present invention. That is, the present invention may be embodied in various forms within the scope of the claims and equivalents thereof, and it is possible for the technician skilled in the art to make various embodiments within the scope of the technical idea of the present invention. .

Claims

A storage unit for storing and accumulating quality information of the object to be processed associated with the state information of the object to be processed and the state information;

A measurement unit for acquiring current state information of the object to be processed in the object processing path;

And a quality predicting unit for extracting a data set corresponding to the current state information based on the data stored in the storage unit and predicting the quality of the object using the data set.
The method according to claim 1,

Wherein the measuring unit comprises:

A first measurement unit for obtaining an exhaust gas temperature distribution of the object to be processed along the extension direction of the path; And

And a second measuring unit for obtaining a temperature distribution in the object to be processed in a direction across the path at an end point of the path.
The method of claim 2,

Wherein the first measuring unit and the second measuring unit comprise:

A quality prediction apparatus for patterning and acquiring state information in the form of column image information.
The method according to claim 1,

Wherein,

And stores cause information about the quality information together.
The method of claim 4,

Wherein,

The quality information, the quality information and the cause information for the quality information are matched and accumulated in the data set.
The method of claim 4,

And a control unit connected to the quality predicting unit,

Wherein,

And controls at least a part of the object processing facility using cause information matching the predicted quality.
The method according to claim 1,

The quality predicting unit,

A quality prediction apparatus for extracting a data set using an artificial intelligence based in-depth neural network and predicting the quality of the object to be processed.
The method of claim 7,

The artificial intelligence includes deep running,

Wherein the depth neural network comprises a convolutional neural network.
A process of accumulating and accumulating the quality information of the object to be processed related to the state information of the object to be processed and the state information;

Processing the object to be processed in the processing path;

Acquiring current state information of the object to be processed in the processing path;

Extracting a data set corresponding to the current state information from the accumulated data and using the data set to predict the quality of the object to be processed;

And outputting the current state information and the predicted quality according to the current state information.
The method of claim 9,

The process of acquiring the status information includes:

A step of patterning and obtaining an exhaust gas temperature distribution of the object to be processed in the form of thermal image information in the direction of extension of the path;

And patterning the temperature distribution in the object to be processed in the form of thermal image information in the direction crossing the path at the end point of the path.
The method of claim 9,

The quality information of the object to be processed associated with the status information and the status information, the cause information of the quality information, and accumulating the matching result into a data set.
The method of claim 9,

The step of extracting the data set includes:

Comparing the current state information with state information in the accumulated data using an artificial intelligence based in-depth neural network;

Selecting state information in the accumulated data corresponding to the current state information, and extracting a data set including the selected state information;

And predicting quality information of the object to be processed included in the extracted data set with the current quality of the object to be processed.
The method of claim 12,

If there is no state information in the accumulated data that matches the current state information,

And extracting a set of data including accumulated state information that is most similar to the current state information.
The method of claim 11,

After the outputting of the predicted quality,

If the predicted quality is output as fluctuation and abnormality, outputting the cause information together;

And controlling at least a part of the object processing facility using the output cause information.
The method of claim 9,

After the process of predicting the quality of the object to be processed,

Acquiring actual quality information of the object to be processed associated with the current state information;

And adding the current state information and the actual quality information to the accumulated data.
The method of claim 9,

Wherein the object to be processed comprises an sintered ore,

Wherein the treatment path includes a sintering zone through which the sludge bed can pass.