KR102059246B1 - System for processing blast furnace sensing data - Google Patents

Abstract

The present invention relates to a furnace sensing data processing system.
The furnace sensing data processing system of one embodiment of the present invention is arranged in m steps (m is an integer of 2 or more) according to the height of the three-dimensional furnace, and n directions (n is 2 for each height step of the three-dimensional furnace). The above sensors); And a furnace sensing data processing apparatus for disposing the sensing data collected from the sensors into an m_n data set based on the position information of the sensors, and converting the m_n data set into a two-dimensional image.

Description

Furnace sensing data processing system {System for processing blast furnace sensing data}

The present invention relates to a furnace sensing data processing system, and more particularly, to a furnace sensing data processing system for aligning data sensed by a plurality of sensors provided in a furnace based on the position of each sensor and converting the data into an image. will be.

A blast furnace is a furnace used to make pig iron from iron ore, also known as a blast furnace. The furnace is divided into a coke oven, a charcoal furnace, and an electric furnace according to what is used as a heat source. Most of the world's pig iron is produced in a coke oven.

The structure of the blast furnace consists of a cylindrical body (furnace) stacked with refractory bricks, a fuel raw material (for example, iron ore, coke), a hot blast furnace, and the like, and a height of approximately 20 to 30 m and a size (ie capacity). Is usually the t number of pig iron produced per day.

The furnace is equipped with a number of sensors (e.g. temperature sensor, pressure sensor) on the furnace body and top for condition monitoring, and the data sensed by these sensors infer the internal condition of the furnace, gas flow distribution and breathability. Here, the overall situation inside the furnace is referred to as yellowing, and the situation of yellowing is judged by sensors (for example, a temperature sensor and a pressure sensor) provided in the furnace. Actions such as blowing actions are taken.

In the case of the prior art, the data sensed by each sensor provided in the furnace is displayed in the form of numerical data on the user (field operator) terminal, and when the user checks this and detects an abnormal condition, the fuel charge action and the blowhole action are performed. And necessary measures were taken.

However, the user has a large difference in the ability to detect abnormal conditions according to the working skill, which significantly affects the productivity of the operation. In the prior art, the data sensed by each sensor disposed in the three-dimensional structure in the furnace is simply a numerical value. Since it is displayed only in data format, the correlation between adjacent sensors is not expressed in the up, down, left, and right sides, and the status of undetected sections between the sensors is not displayed, so the productivity of operation varies greatly according to the user's judgment ability and is unstable. There was a problem of getting lost.

In particular, the temperature or pressure control in the furnace has a great influence on the quality of pig iron, and if such temperature or pressure control is wrong, the pig iron in the furnace is badly damaged.

Korean Unexamined Patent Publication No. 10-2014-0124031 Korean Unexamined Patent Publication No. 10-2017-0110194

The present invention was devised to solve the above problems, and an object of the present invention is to enable the sensing of the furnace to objectively and easily detect the state of the furnace regardless of the user's working skill. It is to provide a data processing system.

Another object of the present invention is to provide a furnace sensing data processing system that allows a user to know the state of the furnace more accurately by processing the sensing data in consideration of the correlation of the sensors provided in the furnace.

It is still another object of the present invention to provide a furnace sensing data processing system and method for aligning and imaging sensing data collected from sensors provided in a furnace so that a model verified to be excellent in AI model learning can be used.

For the above purpose, the furnace sensing data processing system of one embodiment of the present invention is arranged in m stages (m is an integer of 2 or more) according to the height of the three-dimensional furnace, and n number of steps per height step of the three-dimensional furnace. Sensors arranged in a direction (n is an integer of 2 or more); And a furnace sensing data processing apparatus for disposing the sensing data collected from the sensors into an m_n data set based on the position information of the sensors, and converting the m_n data set into a two-dimensional image.

Preferably, the furnace sensing data processing apparatus converts the numerical value of the sensing data collected in the three-dimensional furnace into a two-dimensional image of mⅹn in correspondence with color. In this case, the magnitudes of the numerical values of the sensing data are converted into the two-dimensional images in correspondence to the change in color so that the left and right and upper and lower correlations of the sensing data can be identified from the two-dimensional image.

Preferably, the sensors arranged in n directions for each height step are divided into 360 degrees n equal to 0 degrees, 360 degrees ⅹ 1 / n, 360 degrees ⅹ 2 / n so as to cover all the circumferences of the three-dimensional furnace. , ..., 360 ° (n-1) / n direction.

Advantageously, the furnace sensing data processing device expands data in the m \ n data set to convert an m \ (n + 1) or m \ (n + 2) data set into a two-dimensional image.

Preferably, the furnace sensing data processing apparatus interpolates the converted two-dimensional image to generate an interpolated two-dimensional image.

Preferably, the positional information of the sensors is such that the position of the sensor with respect to the furnace is {0, 1, 2, ..., m-1} height step and {0, 1, 2, ..., n- M} n two-dimensional array in the direction of 1} or mⅹn two-dimensional array representing the position of the sensor relative to the furnace furnace body in the height and angle of the cylindrical coordinate system.

According to the present invention, the numerical data of the sensing data of the three-dimensional structure can be easily implemented as the two-dimensional image by displaying the sensing data of the sensors provided in the furnace in a three-dimensional structure corresponding to the color for each color as a two-dimensional image. Accordingly, the state of the furnace can be easily and objectively detected regardless of the user's working skill, thereby improving the operation efficiency.

In addition, according to the present invention, by sensing the data from the sensors provided in the furnace to generate a data set to expand and process the sensing data in consideration of the correlation between the sensors, accordingly undetected between the sensors The situation of the section (the state of the section between the sensor and the sensor) can also be known, so that the user can know the state of the furnace more accurately.

In addition, according to the present invention, since the sensing data collected from the sensors provided in the furnace is aligned and imaged, it is possible to use a model that has been proven to be excellent in AI model learning, and between the sensors that could not be represented by numerical data. The AI model can be trained on the undetected section state (the state of the section between the sensor and the sensor).

1 is a block diagram of a furnace sensing data processing system according to an embodiment of the present invention.
2 is a diagram illustrating a structure in which a plurality of sensors are provided in a furnace according to an embodiment of the present invention.
3 illustrates a data set generated after aligning a plurality of sensing data in a furnace sensing data processing apparatus according to an embodiment of the present invention.
4 illustrates a data set generated after aligning and expanding a plurality of sensing data in a furnace sensing data processing apparatus according to an embodiment of the present invention.
5 is a diagram illustrating a furnace sensing data processing method according to an embodiment of the present invention.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings and preferred embodiments. For reference, detailed descriptions of well-known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted in the following description.

Furnace according to the present invention is provided with a furnace sensing data processing system that can specifically know the state of the three-dimensional furnace, so that the user can easily detect the state of the furnace.

1 is a block diagram of a furnace sensing data processing system according to an embodiment of the present invention.

Referring to FIG. 1, a furnace sensing data processing system according to an embodiment of the present invention includes a plurality of sensors 100-1, 100-2,. And a furnace sensing data processing device 200 for receiving and processing sensing data (measured value) from 1, 100-2, ..., 100-mn, and is interlocked with the furnace sensing data processing device 200. It may include a deep learning system 300 for performing AI model training.

The plurality of sensors 100-1, 100-2, ..., 100-mn are similarly provided in three-dimensional structures throughout the furnace body in order to detect the state of the furnace (eg, pressure, temperature). It may be implemented in the form of a pressure sensor, a temperature sensor, or the like.

A plurality of sensors (100-1, 100-2, ..., 100-mn) according to an embodiment of the present invention is disposed in m steps (m is an integer of 2 or more) according to the height of the three-dimensional furnace, In addition, it is arranged in n directions (n is an integer of 2 or more) for each height step of the three-dimensional furnace. Here, n directions are divided into equal parts (360 ° / n) by n equals 360 ° to cover all the perimeters of the three-dimensional furnace, that is, 0 directions, 360 °, 1 / n, 360 ° ⅹ It is preferable to set the 2 / n, ..., 360 ° (n-1) / n direction (for example, 0 °, 90 °, 180 °, 270 ° when n = 4). However, the present invention is not limited thereto, and the n directions may be elastically set according to the structure of the furnace.

In this regard, Figure 2 is a diagram illustrating a structure in which a plurality of sensors are provided in the furnace according to an embodiment of the present invention.

In the case of Fig. 1, the three-dimensional furnace is equipped with pressure sensors in 12 steps depending on the height and in 4 directions (0 °, 90 °, 180 °, 270 ° directions) for each height step. There are 72 temperature sensors in four directions, in height and in height steps.

In addition, the values (sensing data) measured by the plurality of sensors 100-1, 100-2,..., 100-mn are transmitted to the furnace sensing data processing apparatus 200, in which case the individual The sensor may be implemented to transmit only measured values (sensing data), and preferably, the individual sensors may be implemented to transmit sensor identification information, sensing time information, and measured values (sensing data) together.

The furnace sensing data processing apparatus 200 processes and sorts sensing data received from the plurality of sensors 100-1, 100-2,..., 100-mn and based on the image, indicates an image representing the current state of the furnace. Create

To this end, the furnace sensing data processing apparatus 200 according to an embodiment of the present invention includes a data receiver 210, a data aligner 220, an image generator 230, an image interpolator 240, and a display unit ( 250) and the like.

The data receiver 210 receives sensing data from the plurality of sensors 100-1, 100-2, ..., 100-mn, respectively. And, if the individual sensor is implemented to transmit only the sensing data, the data receiving unit 210 refers to a sensor for the corresponding sensing data by referring to the receiving path (by which the sensor can be identified) and the receiving time of each sensing data. Generate location information and sensing time information, and if an individual sensor is implemented to transmit sensor identification information, sensing time information, and sensing data together, the sensor location information corresponding to the sensor identification information is generated by referring to a matching table.

For reference, the sensor identification information is for identifying a sensor, for example, may be implemented in the form of a serial number, and the sensor position information may indicate the sensor position in the furnace furnace body, for example, in the form of a two-dimensional array. . Of course, it is also possible to implement sensor identification information and sensor position information in the same manner.

According to a preferred embodiment of the present invention, the sensor position information is based on the height of {0, 1, 2, ..., m-1} and the position of the sensor relative to the furnace furnace body. , n−1} in mⅹn two-dimensional arrays. And, if more accurate state information about the furnace is needed, the sensor position information is applied to the furnace body by applying a cylindrical coordinate system (cylindrical coordinate system) mⅹn two-dimensional array expressing the position of the sensor in height (z) and angle (θ) Is implemented. For reference, when the position of the sensor is expressed by the height and angle applied by the cylindrical coordinate system, the relative position of each sensor in the furnace body can be accurately represented, and as a result, the state at any point of the furnace during image interpolation I can see exactly.

The data aligning unit 220 receives sensor position information, sensing time information, and sensing data for each of the plurality of sensors from the data receiving unit 210, and sorts the plurality of sensing data based on the respective sensor position information.

For example, the data alignment unit 220 extracts sensing data determined to be sensed within the same or predetermined time range with reference to the sensing time information, and sorts the sensing data based on the respective sensor position information. If a plurality of sensors (100-1, 100-2, ..., 100-mn) are provided in the furnace furnace m m height stages and n directions for each height stage, all m 개의 n sensing data Are aligned, resulting in a two-dimensional mⅹn data set.

FIG. 3 illustrates this, with 48 sensing data received from 48 sensors provided in the furnace in 12 height steps and 4 directions (0 °, 90 °, 180 °, 270 °) for each height step. The 12 ⅹ 4 data set generated by sorting is shown.

On the other hand, according to a preferred embodiment of the present invention, in order to compensate for the disconnection of the image displayed at both ends when converting the sensing data of the three-dimensional structure into a two-dimensional image, the data alignment unit 220 is a plurality of sensing data The data set generated by aligning is extended to the left and / or the right side with respect to the direction to generate the extended data set.

Specifically, in one embodiment, the data alignment unit 220 generates a mⅹn data set by arranging the plurality of sensing data in m heights and n directions, and then outward (left or right) with respect to the n directions. Expand one direction to create mⅹ (n + 1) data sets arranged in (n + 1) directions.

For example, when n directions are 0 °, 360 ° ⅹ 1 / n, 360 ° ⅹ 2 / n, ..., 360 ° ⅹ (n-1) / n direction, the data alignment unit 220 is 0 °. Create a hypothetical -360 ° / n direction on the outside (left) of the direction, then copy a sense data column in the 360 ° ⅹ (n-1) / n direction that is substantially equal to the 360 ° / n direction. Add to create an mn (n + 1) data set, or create a virtual 360 ° direction on the outside (right) side of the 360 ° n (n-1) / n direction, then 0 ° substantially equal to the 360 ° direction. Create a mⅹ (n + 1) data set by copying and adding a sense data column in the direction.

In another embodiment, the data aligning unit 220 generates a mⅹn data set by arranging a plurality of sensing data in m heights and n directions, and then sets the mⅹn data sets to both sides (left and right) with respect to the n directions. Direction-by-direction to generate mⅹ (n + 2) data sets arranged in (n + 2) directions.

For example, when n directions are 0 °, 360 ° ⅹ 1 / n, 360 ° ⅹ 2 / n, ..., 360 ° ⅹ (n-1) / n direction, the data alignment unit 220 is 0 °. Create a hypothetical -360 ° / n direction to the left of the direction-then copy and add a sense data column in the 360 ° ⅹ (n-1) / n direction that is substantially the same as the 360 ° / n direction, In addition, by generating a virtual 360 ° direction on the right side of 360 ° ⅹ (n-1) / n direction, mⅹ (n + is added by copying and adding a column of sensing data in the 0 ° direction that is substantially the same as the 360 ° direction. 2) Create a data set.

Fig. 4 illustrates this, by arranging 48 sensing data received from 48 sensors equipped with 12 heights and 4 directions (0 °, 90 °, 180 °, and 270 °) in the furnace furnace. After generating, expand in one direction on both sides with respect to four directions (0 °, 90 °, 180 °, 270 °) and then expand the six directions (-90 °, 0 °, 90 °, 180 °, 270 °, Create a 12x6 data set arranged in 360 °). Here, the newly added sensing data for the -90 ° and 360 ° directions is the same as the 270 ° (= -90 ° + 360 °) and 0 ° (= 360 °-360 °) directions, respectively. Created by copying data.

As such, in a preferred embodiment of the present invention, by adding continuous and highly relevant sensing data for the sorted mⅹn data set and extending it to the mn (n + 1) or mⅹ (n + 2) data set, 3 Even if the sensing data of the dimensional structure is converted into a two-dimensional image, since the image is extended to the outside and a continuous image is added, the association between the images displayed at both ends is compensated for.

The image generator 230 generates a two-dimensional image by mapping the numerical values of the two-dimensional data sets generated by the data aligning unit 220 to colors for each element. In this case, the two-dimensional image is generated by corresponding the magnitude of the numerical value of the sensing data to the change of the color so that the left and right relations and the upper and lower relations of the sensing data can be confirmed.

For example, the image generator 230 may generate a two-dimensional image by converting a numerical value of sensing data into a corresponding color for each element of the mⅹn data set generated by the data sorter 220 and converting the numerical value of the sensing data into a corresponding color. (See FIG. 5A).

In this case, the magnitude of the numerical value is expressed in response to the change in color. In FIG. 5, as the numerical value gradually increases from the small value to the large value, the color gradually changes to navy blue, blue, green, yellow green, and yellow to correspond.

If the data set is expanded by the data sorter 220, the image generator 230 may be configured to each of the mⅹ (n + 1) or mⅹ (n + 2) data sets extended by the data sorter 220. The numerical value of the sensing data is converted to the corresponding color for each element and displayed in the mⅹ (n + 1) or mⅹ (n + 2) zone to generate a two-dimensional image (see FIG. 5B).

The image interpolator 240 generates an interpolated two-dimensional image by interpolating the two-dimensional image generated by the image generator 230.

For example, the image interpolator 240 interpolates a two-dimensional image generated by the image generator 230 by performing a known image interpolation method (eg, Nearest Neighbor, Bilinear, Bicubic), and interpolates the color change smoothly. Generate a two-dimensional image (see FIG. 5C).

In addition, the display unit 250 displays the interpolated two-dimensional image generated by the image storage unit 240 so that the user can see it.

With the above-described configuration, according to the present invention, since numerical analysis of the sensing data of the three-dimensional structure is implemented as a two-dimensional image, it is possible to easily determine whether there is an abnormality in the furnace by discriminating the color change of up, down, left and right. Can be. That is, the sensing data of the upper and lower sensors can be distinguished by color, so that it is easy to determine whether there is an abnormality in the furnace for a specific height in the same direction (for example, the larger the difference in the color of the image adjacent to the upper and lower sides, the more abnormal is determined). Similarly, the sensing data of the left and right sensors can be distinguished by color, so that it is easy to determine whether there is an abnormality in the furnace in a specific direction of the same height (for example, the larger the color difference between the left and right adjacent images, the more abnormal there is). .

In order to compensate for the disconnection of the image displayed at both ends when the sensing data of the 3D structure is converted into the 2D image, the sorted mⅹn data set is mⅹ (n + 1) or mⅹ (n + 2) data. Since it is extended to a set, even if the 3D sensing value is converted into a 2D image, it is possible to accurately determine whether there is an abnormality in the furnace without substantially different from the result of the 3D implementation.

Meanwhile, the deep learning system 300 performs artificial intelligence model training based on the image transmitted from the furnace sensing data processing apparatus 200 (specifically, the image interpolator 240).

For reference, in the deep learning system 300, an algorithm (eg, an image CNN (Convolutional Neural Network) algorithm) that is effective when learning the image data has already been proved in various cases, and the present invention enables image data learning. Therefore, it is possible to learn the AI model about the situation of the undetected section (the state of the section between the sensor) between the sensors that could not be represented in the numerical or line graph.

Although the present invention has been described in detail with reference to preferred embodiments, it will be apparent to those skilled in the art that the present invention may be implemented in various other specific forms without changing the technical spirit or essential features of the present invention. The described embodiments are to be understood in all respects as illustrative and not restrictive.

In addition, the scope of the present invention is specified by the appended claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and equivalent concepts are included in the scope of the present invention. Should be interpreted as

Claims (9)

Sensors arranged in m steps (m is an integer of 2 or more) according to the height of the 3D furnace, and arranged in n directions (n is an integer of 2 or more) for each height step of the 3D furnace; And
A furnace sensing data processing apparatus for disposing sensing data collected from the sensors into an mⅹn data set based on the position information of the sensors, and converting the mⅹn data set into a two-dimensional image;
Sensors arranged in n directions for each height step are divided into 360 degrees n to cover the circumference of the three-dimensional furnace, and 0 °, 360 ° ⅹ 1 / n, 360 ° ⅹ 2 / n, ..., 360 ° Ⅹ is arranged in the (n-1) / n direction,
The furnace sensing data processing apparatus expands data in the mⅹn data set to convert mⅹ (n + 1) data set into a two-dimensional image,
The mⅹ (n + 1) data set adds 360 ° ⅹ (n-1) / n direction sensing data outside the 0ⅹ direction sensing data in the mⅹn data set or 360 ° ⅹ (n-1) / Furnace sensing data processing system, characterized in that the expansion by adding the sensing data in the 0 ° direction outside the sensing data in the n direction. The method of claim 1,
The furnace sensing data processing apparatus converts a numerical value of the sensing data collected in the three-dimensional furnace into a two-dimensional image corresponding to a color. The method of claim 2,
The furnace sensing data processing apparatus converts the magnitude of the numerical value of the sensing data into the two-dimensional image in response to a change in color so that the left and right relations and the vertical relations of the sensing data can be identified from the two-dimensional image. Furnace sensing data processing system, characterized in that. delete delete The method of claim 1,
The furnace sensing data processing apparatus expands data in the mⅹn data set to convert mⅹ (n + 2) data set into a two-dimensional image,
The mⅹ (n + 2) data set adds 360 ° ⅹ (n-1) / n direction sensing data to the mⅹn data set outside of the 0 ° direction sensing data, and also adds 360 ° ⅹ (n-1) / Furnace sensing data processing system, characterized in that the expansion by adding the sensing data in the 0 ° direction outside the sensing data in the n direction. The method according to any one of claims 1 to 3,
The furnace sensing data processing system interpolates the converted two-dimensional image to generate an interpolated two-dimensional image. The method according to any one of claims 1 to 3,
The positional information of the sensors indicates the position of the sensor with respect to the furnace in the height step of {0, 1, 2, ..., m-1} and the direction of {0, 1, 2, ..., n-1}. 2. A furnace sensing data processing system, characterized in that it is implemented by a mⅹn two-dimensional array or m 또는 n two-dimensional array expressing the position of the sensor relative to the furnace furnace body by the height and angle of the cylindrical coordinate system. A furnace body having a three-dimensional shape; And
A furnace comprising a furnace sensing data processing system according to any one of claims 1 to 3.

Images