KR960001823B1 - Powder cokes grain size distribution inspecting system - Google Patents

Abstract

The continuously measuring apparatus improves the quality and the manufacturing cost of the cokes powders by controlling mixing of raw materials optimally. The apparatus comprises: feeder(1) driven by a motor(2); a drier (4) with hot air injected by air heater(3); a hopper(5) for transferring cokes powders to the electro-disperse vibrator(6); and a black box(9) with lance and a camera(7).

Description

Continuous measuring method of powder coke size distribution and its device

1 is a schematic diagram of a particle size distribution continuous measurement apparatus according to an embodiment of the present invention.

2 is a detailed view of an injection unit continuous sampling apparatus in the present invention.

Figure 3 is a detailed view of the injection body imaging device in the device of the present invention.

4 is a detailed configuration diagram of an image processing system for particle size measurement in the apparatus of the present invention.

5 is a schematic representation of a typical binarized particle image.

6 is a flow chart of particle size distribution measurement according to the present invention.

7 is a view showing the characteristics of sintered ore according to the coke particle size.

8 and 9 are schematic views of a conventional image particle size distribution measuring apparatus.

* Explanation of symbols for main parts of the drawings

1: Screw Feeder 2: Motor

3: air heating device 4: sample heating and drying device

6: Electromagnetic Dispersion Vibration Device 7: CCD Camera

8, 8 ': Window 9: Dark Box

10 halogen lamp 11: translucent plate

12 cooling fan 13 inner cylinder

14: lighting box 15: computer

16: frame grabber board 17: frame processor board

18: Color Video Monitor 19: Color System Monitor

27: coke bean 28: CFW

29: image processing unit

The present invention relates to powder coke sintering for sintering iron ore in steelmaking in steelworks, and more particularly, to a method for continuously measuring and measuring a particle size distribution of water-containing powder coke in a blended raw material of a sintering process.

The powdered coke acts as a main heat source in the production of sintered ore, and even when the same amount of powdered coke is blended, the burning speed, the maximum reaching temperature, and the high temperature holding time vary depending on the particle size distribution of the powdered coke, as shown in FIG. It has a great influence on the room temperature strength (TI), reduction differentiation strength (RDI) and reduction rate (RI) of the. These powdered cokes vary greatly in the shape and size of the coke raw materials used for grinding, so the particle size distribution of the powdered coke needs to be continuously measured before the powdered coke is used in the grinding process or the blended raw materials. .

Conventionally, the coke that has been transported has been sampled at regular time intervals (usually 4-8 hours), dried, and then sifted (measured by weight) to measure particle size distribution, but the measurement takes only a few hours. The variation of the particle size distribution between the cycles was so severe that the measured values could not be applied to the grinding or blending process. Recently, continuous particle size distribution measurement methods have been attempted, and representative methods include laser diffraction and image processing. The method using the laser diffraction is applied to the particle size measurement of the unit of several hundred μm or less, but requires expensive equipment cost.

In addition, as shown in FIGS. 8 and 9, the image processing method measures the particle size falling from the conveyor through a camera or a halogen lamp and a linear sensor, and then displays it on a CRT screen by processing it with a computer. Particles containing about 10% moisture were not able to measure particle size distribution because they can be applied only to the measurement of particles having a size of several mm to several tens of mm or more.

The present invention is a method that can continuously measure the particle size distribution of powdered coke in the image processing method that contains a large amount of moisture and the particle size distribution is dried from several tens of micrometers to several mm dry coke and separated and dispersed between the particles and Suggest a device.

Features of the present invention include the steps of sampling a certain amount of bulk coke that is being transported or falling, drying the sampling coke, dispersing it, obtaining a momentary still image of the particles in the fall state, and characterizing the particles from a still picture. Obtaining a quantity, and performing continuous repetition of calculating and estimating a particle size distribution from image processing data and performing statistical processing.

Hereinafter, the present invention will be described in detail.

In the present invention, the particle to be measured is not limited to any particular three-dimensional shape, and can be widely applied to powdered coke, iron ore, limestone, sintered ore, etc. containing a large amount of water.

1 to 3 are explanatory views of the apparatus of the present invention, when the water-containing powdered coke A supplied from the coke bin 27 is conveyed through a continuous feeding welghter (CFW) 28. The sample is installed to be sampled by the heating and drying apparatus (4). The screw feeder (1) is installed to move by the motor (2), the sample heating and drying device 4 is installed so that the air heated in the air heating device (3) is injected. The powdered coke (A ') dried in the sample heating and drying device (4) is transferred to the electron-dispersion vibrator (6) through the hopper 5, and the dry powder coke (A' falling from the electron-dispersion vibrator). ) Is placed to pass between the light box 14 and the dark box (9).

The dark box 9 includes a window 8 'and a camera 7. The lighting box 14 is composed of an inner cylinder 13 including a cooling fan 12, a halogen lamp 10, a translucent plate 11, a window (8). On the other hand, the video signal output from the camera 7 is provided to the image processor 29. The image processor 29 has a hardware configuration as shown in FIG. The video signal output from the CCD camera 7 is configured to be output to the color video monitor 18 via the frame grabber board 16, and the frame processor board 17, the memory 31, the mouse 33, the keyboard ( 34) The input is processed by the general-purpose computer 15 and configured to be output to the color system monitor 19 and the printer 32.

Accordingly, in the present invention, the open end of the screw feeder 1 is embedded in the flow of the injection body A carried on the conveyor belt or poured from the grinder, and the screw feeder 1 is connected to the reduction motor 2. When rotating at a constant speed, a certain amount of sieve sample is continuously collected from the flow velocity A of the sieve of the measurement object. The sampled sieve is first indirectly heated by the heated air of the sample heater 4 installed in the double wall on the outer wall of the screw feeder 1, and passed through the heated air to the center of the screw feeder 1. To remove the moisture contained.

The dried coke coke (A ') falls on the diaphragm of the electromagnetic vibration dispersing device 6 and is separated and dispersed between the particles and freely falls at the end of the diaphragm.

The particle A 'falling from the electromagnetic vibration dispersing device 6 passes vertically between two glass windows 8 and 8' whose surface is hardened (deposited by 10 m A1203). In this case, the CCD camera 7 is photographed at a shutter speed of 1 / 2,000-1 / 10,000 seconds and transmitted as a TV video signal. In addition, a glass window 8 is installed on the lens side of the CCD camera 7 and the entire CCD camera 7 is sealed in the dust and dark box 9 to reduce the influence of external light intensity change and to protect it from dust. In addition, since the sensitivity of the CCD camera 7 is inversely proportional to the shutter speed and proportional to the illuminance, illumination corresponding to the shutter speed is required.

In the present invention, the front of the lens of the CCD camera (7), that is, the opposite side through which the injection unit (A ') is given a strong illumination by the halogen lamp 10, and uniformity of the brightness and heat generated from the halogen lamp (10) In order to block, the translucent plate 11 installed in the intermediate position is sealed in the inner cylinder 13 and the glass window 8, and is prevented from being contaminated. In addition, since the halogen lamp 10 generates several hundred W of heat, the cooling fan 12 is installed at the rear end to cool it, and the lighting box 14 surrounding the inner cylinder 13 is installed to increase the cooling effect. Create a large space on the side 8 to allow the coolant to pass through.

The still image signal of the moving object captured by the CCD camera 7 and transmitted is analyzed by the image processing unit 29. The analog video signal transmitted from the CCD camera 7 is quantized (digitized) by the frame grabber 16 and stored as a digital video in a memory buffer of 512 x 512 x 8 bits. In this case, the image is expressed as 8 bits of the grayscale of the image, that is, 256 gray levels.

The particle information captured in the digital image is processed by the frame processor 17 by arithmetic or logical operation, convolution, histogram and the like. These image processing boards 16 and 17 are directly installed in an I / O slot of the general purpose personal computer 15, and the general purpose personal computer 15 is provided with the image processing boards 16 and 17. Control is performed to perform mathematical equation statistical processing, data storage, etc. of algorithms and measured data that the image processing board cannot process. In addition, the color video monitor 18 displays an input analog image or an output digital image, and the color system monitor 19 displays a histogram or a measurement result.

6 is a particle size distribution measurement program by the image processing unit 29 of FIG. 4, in which an image signal transmitted from the CCD camera 7 is subjected to image sampling (20), light image processing (21), and binary image processing (22). The image processing is performed in the order of binary image processing (23), measurement (24), statistical analysis (25), and result display (26). The transmitted video signal is interlaced by scanning the first scan line after the first scan after the first scan in a raster scan method such as a home TV. That is, since one frame is formed of two odd and even scan lines (Odd and Even Field) and 30 screens are transmitted in one second, the time between fields takes 1/60 second.

Since the particle | grain A 'falls freely, it moves about 23 nm in 1/60 second at the position of 10 cm of fall distances. In the present invention, a special technique is used to solve this problem. In other words, one of the draft and even fields was ignored, and one field data was duplicated to have the same information as the other field data. This method has the disadvantage that ± error occurs by one pixel in the vertical length direction of the particle and 50% of the information on one pixel disappears.

However, this method is simple and can be solved by measuring the particle size in the horizontal direction because the object of the present invention is to measure the distribution of particle size.

The image information thus obtained is subjected to the shaded image processing 21 using a linear transformation or a spatial filter technique to correct and improve the change of illumination, lens contamination, chipping of particles, and image noise. The signal is binarized 22. By specifying the 256-gray level, which is binarized, the dark part is divided into "black" and the bright part is divided into "white", that is, the black and white (0 and 1) are binarized (Binaeyzation). Is called processing. The binarized image is repeatedly expanded or shrunk in four areas of eight areas or a combination thereof in the case where two particles are connected by a thin line, overlapped, or one particle is divided into two. In other words, it is properly cut or connected by the binary image processing 23.

The method of metrology 24 uses an area labeling algorithm for each particle. The labeling refers to the process of numbering and processing the connecting areas in order for a binary image. A set of black pixels connected to each other is called a single "black" area. The labeling method includes a run length measurement or a contour extraction method. In the present invention, the algorithm used is a contour extraction method. Since the particles are represented by the black region, the contour of the black region, that is, the boundary line with the white region is extracted. The contour extraction method uses a 3x3 mask or a 2x2 mask. The 2 × 2 mask 4-connected line extraction method used in the present invention is as follows.

5 is a typical particle image, where the lower left (X k) of the 2x2 mask such as ⓐ is positioned at the coordinates (X13, Y1) where the scan line first meets "1", and the direction of travel is indicated by the arrow (↑) direction.

If ⓐ is a white pixel (even though b is a black pixel, b is ignored because it corresponds to 8 near points rather than 4 near points), rotate 90 ° counterclockwise around the current contour point (Xk) as shown in ⓑ. do.

If ⓐ is a black pixel and ⓑ is a white pixel, move the outline coordinates one by one in the advancing direction.

If ⓐ and ⓑ are all black pixels, rotate the current direction 90 ° clockwise as shown in picture ⓓ and move the contour point to the right corner.

In this way, if the particle boundary of the black region is circumscribed while the following Xk + 1 and Yk + 1 are determined according to the black and white states of ⓐ and ⓑ in the advancing direction, the X and Y coordinate values of each boundary point are suppressed. The particle size and area are obtained from these X and Y coordinate values. The labeled black region changes the color value of the pixels to achieve labeling overlap. When this operation is repeated for all black region particles on the screen, when the scanning line reaches the final line of the screen, the number of particles on one screen, X, Y coordinate values for each particle, number of pixels, and the like are obtained.

The statistical analysis (25) includes calculating particle size, area, and surrounding length from the particle number and particle information thus obtained, and calculating the number of particles based on correlation and classification criteria. In order to convert the particle size, area, etc. into actual values, the length per pixel and the area per pixel are multiplied by the magnification of the camera. In addition, in order to know the weight percent (particle size distribution) and the like of each classification, a correlation between the measured particle count and the actual description is obtained.

After the statistical analysis processing, the particle size distribution information is displayed on the screen or stored in the storage, and the process is returned to the image sampling 20 which is the first order and the processing is repeated.

The following describes an embodiment of the present invention. The CFW 28 is a device for continuously supplying a blended raw material of sintered ore. In the case of powdered coke, the CFW conveyor supplies about 15 Ton / hr of coke while proceeding at a speed of 8 m / min. The tip of the screw feeder (1) of the branch cord (A) and continuous quantitative sampling apparatus of the present invention is installed in front of the conveying direction of the CFW (28), and the opening of the screw feeder (1) is the coke (A) just before the dropping. I get buried in).

Since the powdered coke A supplied from the conveyor of the CFW 28 contains about 10% of water, most of the coke powder adheres to each other and condenses to form a large mass.

Therefore, since the particles are not separated and dispersed unless the moisture is dried, in the present invention, the sampled coke was dried using air heated at 120 ° C. Table 1 shows the change in moisture of the powdered coke before and after the sampling device of one sample continuously sampled at 5 minute intervals.

[Table 1. Moisture comparison before and after sampling drying]

The dried powder coke sample (A ') is dispersed and falls while passing through the electromagnetic vibration disperser (6). The falling coke particles change the particle A 'into a planar still image and then send a TV video signal to the social cab. Coke granularity information is transmitted 30 frames per second and it takes more than a few seconds to measure one frame depending on the number of particles. In addition, the number of particles measured in one frame is only a few hundred, and only a fraction of the estimated particles are actually captured. And since the particle size distribution is very wide from tens of micrometers to several millimeters, it is necessary to increase the number of measured particles to reduce the measurement error.

In the present invention, the granularity measurement data of several hundred frames are accumulated for a predetermined time and the average value thereof is output. The particle information thus measured is a value excluding most values of several hundred μm or less as shown in Table 2.

[Table 2. Particle Size Distribution and Particle Number of Bunkosk]

* The number of image processing particles is the average number of particles of 80 frames for 10 minutes.

Therefore, the present invention examines the correlation between particles of 0.25 mm or more that can be relatively measured and particles of 0.25 mm or less that are difficult to measure, and shows particles of 0.25 mm or less that are difficult to measure from a particle size distribution of 0.5-8 mm that can be recalled. The particle size distribution can be estimated from

(Rare model) SDi = Ro + SIRli + ‥‥‥‥‥‥ SnRni

Where SDi is the particle size distribution (% by weight)

i is +.25, +.125, -125 mm

S1-Sn is a weight percent of +8 mm to +0.5 mm

R1i to Rni are the correlation coefficients for the i particle size distribution

Table 3 shows the results of the particle size analysis obtained from the image processing and the rare model equation, and shows a very high correlation with the body analysis value of 0.8 or more.

[Table 3. Image Analysis Results]

S: sieve analysis value (wt%)

I: chemical conversion value (wt%)

E: Estimated value by wt% (wt%)

As described above, the particle size distribution of a particle having a particle size distribution of several tens of micrometers to several millimeters containing a large amount of water, which has not been continuously measured by sieve analysis or image processing, can be obtained by the particle size analysis method and apparatus of the present invention. You can measure continuously. Such continuously measured particle size distribution analysis data can be fed back to the granular crushing process or the feed material control system and automated to enable optimal control, thereby greatly contributing to quality improvement and cost reduction.

Claims (2)

The powdered coke during mass transportation is continuously sampled with a screw feeder, and the sample is supplied with air heated by a heat and drying device, dried, dispersed by an electronic vibration dispersion device, and photographed with a CCD camera while dropping it. Select one of the even and odd fields of the photographed image data and copy it to another field to extract the instantaneous still image data, and calculate the particle size distribution of particles having a diameter of 0.25 mm or more by shaded the image data by a spatial porter technique. And the particle size distribution of the particles of 0.25 mm or less is SDi = Ro + S1R1i +... SnRni where SDi is the particle size distribution (% by weight), i: 0.25, 0.125, -0.125 mm, S1 to Sn: weight% of +8 mm to +0.5 mm, and R1 to Rni: i particle size distribution. A continuous measuring method of powdered coke particle size distribution, characterized by calculating. In the step of measuring the particle size distribution of the powdered coke supplied from the coke bin to the conveyor belt, a screw feeder for continuously quantitatively sampling the powdered coke supplied from the conveyor belt, and the powdered coke supplied by the screw feeder A sample heating and drying apparatus for drying with heated air, an electromagnetic vibration disperser for separating, dispersing and dropping the dried powdered coke supplied from the sample heating and drying apparatus, and an injection body of the buncosk falling from the electronic automatic disperser. CCD camera photographed by a camera, a halogen lamp installed in front of the CCD camera and illuminating the moving object passing in front of the CCD camera 7, and the bunk coke includes an image processor for receiving and processing powdered coke image data from a photographing unit. Continuous measuring apparatus for the powder coke particle size distribution characterized in that.