JP4244772B2 - Flowing glass abnormality monitoring device - Google Patents

Description

  The present invention relates to a falling glass monitoring device that flows down from a melting furnace to a canister when vitrifying high-level radioactive liquid waste, and more particularly to a falling glass abnormality monitoring device for monitoring the falling glass by image processing. is there.

  Conventionally, in a nuclear reactor facility, as shown in FIG. 6, waste liquid containing high-level radioactive material generated is mixed with molten glass by a melting furnace 21, and glass containing radioactive waste is caused to flow down into a canister 22. inject. The glass was cooled and solidified and sealed in the canister 22 so that radioactive waste generated in the nuclear reactor facility could be stored in the storage facility.

  The canister 22 into which the glass has been poured up to the specified amount is moved on the floor 30 by the carriage 23 and is carried into the storage facility.

  The abnormality of the falling glass 25 flowing down from the melting furnace 21 of the vitrification facility as shown in the figure has been judged by the operator observing the flowing state visually or by monitoring with a monitor.

  There are various causes of the flow-down abnormality of the flow-down glass 25. As a typical cause of abnormality, the heating in the melting tank 21 becomes uneven due to abnormality of the main electrode 61, the bottom electrode 62, the furnace bottom auxiliary electrode 63, etc. of the melting furnace lower part 21a shown in FIG. It is thought that it is.

  Further, the metal particles contained in the radioactive waste mixed in the melting furnace 21 absorb the high frequency current generated by the main electrode 61 and the like, and the heating in the melting furnace 21 may not be sufficiently performed. .

  Such an abnormality such as insufficient heating is a phenomenon in which when the falling glass 25 flows down from the falling nozzle 66, the glass flow is swayed, or the viscosity of the flowing glass is increased to cause swelling. appear. An abnormal monitoring method has been adopted in which an experienced operator observes these phenomena and estimates the cause of the abnormality from the abnormal phenomenon.

  Moreover, a molten glass flow monitoring device using an optical fiber has been proposed (see Patent Documents 1 and 2).

Japanese Patent Laid-Open No. 9-281294 Japanese Patent Laid-Open No. 9-243797

  However, the conventional method relying on visual observation has a problem that skill is required for the judgment of the operator. In addition, there is a problem that the judgment varies due to individual differences.

  In addition, in the case of the molten glass flow monitoring device, the abnormality is output only when the glass flow becomes abnormal, but the situation that is a precursor to the occurrence of the abnormality and the trend until the abnormality occurs are recorded. There was a problem that the abnormal phenomenon could not be predicted or was insufficient.

  Therefore, an object of the present invention is to provide a falling glass abnormality monitoring device capable of collecting a flowing state from image data and predicting an abnormality sign and determining an abnormality.

The present invention has been made in order to achieve the above object, a first aspect of the present invention, the molten glass obtained by mixing radioactive liquid waste in the melting furnace, allowed to flow down as a falling glass from falling nozzle, the falling glass When the canister is filled, in the falling glass abnormality monitoring device for detecting the abnormality of the falling glass flowing down from the melting furnace , a camera for imaging the flowing glass at regular intervals , and sequentially at regular intervals. The captured image of the falling glass is binarized, and the binarized image is scanned in the horizontal direction to detect the left edge portion and the right edge portion of the falling glass, and the detected left edge portion and right edge portion Is a falling glass abnormality monitoring device provided with a data processing device that calculates a center position from the image and detects a fluctuation of the falling glass from a change in the center position calculated for each sequentially captured image .

In a second aspect of the invention, the left edge portion is detected by scanning in the horizontal scanning direction, and the left edge portion is detected, and the right edge portion is detected in the direction opposite to the horizontal scanning direction. To detect the right edge portion.

According to a third aspect of the present invention, the detection of the left edge portion starts scanning in the scanning direction from a position separated from the left edge portion detected last time by a predetermined distance to the left, and the detection of the right edge portion detects the right edge detected last time. Scanning is started in the direction opposite to the scanning direction from a position a predetermined distance away from the part to the right.

According to a fourth aspect of the present invention, the detection of the shaking is detected when a horizontal difference between the center position of the falling nozzle and the center position calculated from the left edge portion and the right edge portion exceeds a threshold value. .

According to a fifth aspect of the present invention, the detection of the shaking is performed by detecting each left edge portion and each right edge portion by two scanning lines having different vertical positions by the scanning, and detecting each detected left edge portion. The center position is calculated from each of the right edge portions, and the shaking is detected when the horizontal difference between the calculated upper and lower center positions exceeds a threshold value.

In a sixth aspect of the invention, the shake is detected when the horizontal difference between the past center position and the latest center position exceeds a threshold value.

In a seventh aspect of the invention, the data processing device counts and stores the number of detected shaking times.

  According to the present invention, it is possible to collect the flow-down state from image data and to exhibit an excellent effect of enabling prediction of an abnormality sign and determination of abnormality.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described with reference to the accompanying drawings.

  FIG. 1 shows a falling glass abnormality monitoring apparatus and its installation state according to a preferred embodiment of the present invention.

As shown in the figure, in the nuclear reactor facility, a melting furnace 21 for mixing and melting radioactive waste with glass, and a canister 22 for pouring the molten glass melted in the melting furnace 21 for injection and filling and storage. And a dolly 23 for conveying the canister 22 and a falling glass abnormality monitoring device 1 for monitoring the condition of the falling glass 25 injected from the melting furnace 21 into the canister 22 and detecting an abnormality.

  Here, the falling glass 25 represents the state of the molten glass that flows down immediately before the molten glass melted in the melting furnace 21 flows down from the melting furnace lower part 21a and is poured into the canister 22, and the falling glass 25 is It shows that it is the object monitored by the falling glass abnormality monitoring device 1.

  The melting furnace 21 is a facility for mixing the high-level radioactive liquid waste generated in the nuclear reactor facility with the molten glass obtained by melting the raw glass so as to flow down as the falling glass 25.

  The canister 22 is made of metal for injecting a falling glass 25 formed by melting in the melting furnace 21 and mixing radioactive waste liquid, sealing the flowing molten glass, and solidifying and storing the radioactive waste. The container. The canister 22 has a cylindrical shape with a diameter of about 40 cm, a height of about 100 cm, and a weight of about 400 to 500 kg.

  The carriage 23 is a carrier for moving the canister 22 on the floor 30 and transporting it to a storage facility when the falling glass 25 injected into the canister 22 reaches a specified amount.

  The falling glass anomaly monitoring device 1 has a camera 3 installed facing the falling glass 25 so as to take an image of the entire state of the flowing glass, and image-processed data obtained by performing image processing on the image captured by the camera 3. The data processing device 2 for detecting the shaking of the falling glass 25 from the stored and stored data, the camera control device 5 for controlling the photographing of the camera 3, the data processing device 2 and the camera control device 5 are remotely controlled. And a remote control system 6 for operation, which are connected by cables 4, 7 and 8.

  The falling glass abnormality monitoring device 1 is a monitoring device for detecting an abnormality of the falling glass 25 flowing down from the melting furnace 21 when the canting glass 22 is filled with the falling glass 25 from the melting furnace 21. It has a function of detecting the shake of the falling glass 25 by a data processing device by image processing of the captured image.

  The falling glass abnormality monitoring device 1 captures an image of the falling glass 25 injected into the canister 22 from the melting furnace 21 in the nuclear reactor facility, and analyzes the captured state image of the falling glass 25 to When an abnormality such as a shake occurs in the state, the abnormality can be detected and a sign of abnormality can be detected.

  The data processing device 2 is a data processing unit 11 for processing image data and the like, a control signal input board 12 for inputting a control signal for controlling the camera 3 to the camera control device 5, and photographing by the camera 3. An image processing board 13 for digitally processing the processed image, image analysis software 14 for digitally processing the image by the image processing board 13, and no power source for backing up data of the data processing device 2 in the event of a power failure In order to print out the processing data and the like, the power storage unit 15, the data storage unit 16 for storing the data processed by the data processing unit 11, the display 17 for displaying the operation state and the state of the processing data, etc. The printer 18 is configured.

  The data processing device 2 binarizes the images of the falling glass 25 sequentially captured by the camera 3 at regular intervals, and scans the processed binary image in a horizontal direction with respect to the flowing direction of the flowing glass. This is a processing device for detecting the left edge portion and the right edge portion of the falling glass 25 and detecting the presence or absence of an abnormality in the flowing state of the glass from these data.

  The uninterruptible power supply unit 15 detects a power failure when the power supplied to the falling glass abnormality monitoring device 1 is stopped for some reason. The uninterruptible power supply unit 15 that has detected a power failure is a backup power supply for switching from an external power supply to a built-in battery and supplying power to the data processing device 2 instead of the external power supply. When the power is normally supplied from the outside, the uninterruptible power supply unit 15 supplies power to each member of the falling glass abnormality monitoring device 1 as it is, and charges a battery (not shown) incorporated therein at the same time.

  The data processing unit 11 is an arithmetic processing unit that performs arithmetic processing on the data image-processed by the image processing board 13, determines whether or not the falling glass shakes, and outputs the calculation result to the data storage unit 16.

  The data storage unit 16 is a memory unit for sequentially storing and storing image processing and calculation results processed and analyzed by the image processing board 13 and the data processing unit 11.

  The display 17 is a display for displaying an operation state of the falling glass abnormality monitoring device 1, a state of processing data, and the like.

  The printer 18 is a printer for printing out image processing data, the presence / absence of shaking indicating an abnormality of the falling glass, the number of shaking, the state of the falling glass abnormality monitoring device 1, and the like.

  The camera 3 is a camera that is installed to photograph the state of the glass injected from the melting furnace 21 into the canister 22, and the glass flowing state is displayed in front of the flowing position so that the glass shakes and the like are captured on the screen. Or install it close to it. As the camera 3, for example, a video camera using a CCD image pickup tube, a near infrared camera excellent in temperature detection, or the like is used.

  Although one camera 3 is shown in the figure for explanation, it is better to install two or more cameras on the circumference centering on the longitudinal direction of the falling glass 25 and photograph the flowing glass 25. The state of the glass 25 can be better monitored, and the shaking of the falling glass 25 can be captured in multiple dimensions.

  The camera control device 5 is a control device for receiving a control signal supplied from the control signal input board 12 in the data processing device 2 and controlling photographing of the camera 3.

The cables 4 and 8 transmit images captured by the camera 3 to the data processing device 2 via the camera control device 5, and control signals supplied from the control signal input board 12 of the data processing unit 2 are transmitted to the camera control device 5. And a signal line for transmission to the camera 3.

  The remote operation system 6 is an operation system for operating the falling glass abnormality monitoring device 1 from a remote monitoring room or the like to acquire and operate data processed by the data processing unit 2.

  The cable 7 is a transmission line for transmitting the data processed by the data processing device 2 to the remote operation system 6 and transmitting the operation of the remote operation system 6 to the data processing device 2. Used.

  Next, the operation | movement outline | summary of the falling glass abnormality monitoring apparatus 1 is demonstrated.

  The falling glass 25 mixed and melted with the radioactive waste liquid in the melting furnace 21 is injected and filled into the canister 22 mounted on the carriage 23, and the state of the falling glass 25 is imaged by the camera 3.

  An image indicating the state of the falling glass 25 captured by the camera 3 is input to the image processing board 13 via the cable 4, the camera control device 5, and the cable 8.

  The image input to the image processing board 13 is smoothed and binarized by the image processing board 13 and subjected to image processing such as contraction and expansion processing for removing noise from the image, and then the data processing unit 11. Is output.

  FIG. 2 is a flowchart showing a procedure for processing the image (shown in FIG. 3) of the falling glass 25 taken by the camera 3 with the image processing board 13 and the data processing unit 11.

  When image data is input to the image processing board 13, digital processing for image processing is started in step 31 shown in the flowchart.

  In the image processing, an image processing range is set for detecting the edge of the glass flowing down from the photographed image as shown in FIG. 3, and the area where the falling glass 25 is reflected and other areas are smoothed by smoothing the screen. Each region is processed as a continuous plane, binarized by a set threshold value, and the left and right edge portions of the falling glass 25 are sharpened by shrinkage and expansion processing, and the left and right edge portions shown in FIG. It is a series of image filtering processing that facilitates detection processing.

  In the image processing board 13, a range where the falling glass 25 is captured is set in order to process the image from the entire image (shown in FIG. 3) captured by the camera 3 (step 32).

  The setting of the image range for detecting the position of the falling glass 25 is set in consideration of the nature of the target position information and the processing time of the image processing board 13.

  The range is set so that the state of the molten glass flowing down from the melting furnace lower part 21a, particularly the range in which the shaking or swelling of the glass can be sufficiently detected, is selected, so that the center position of the flowing nozzle 66 in the melting furnace lower part 21a becomes clear. It is better to set the range.

  When setting the image processing range, the image is captured so that the vertical and horizontal directions of the image coincide with the actual vertical and horizontal directions where the glass flows down.

  If it is determined that the camera 3 is tilted by this correction, the determination result is sent to the control signal input board 12 and transmitted to the camera control device 5 via the cable 8. The camera control device 5 A control signal for correcting the posture of the camera 3 is issued to the camera 3. The camera 3 that has received the control signal for correcting the posture corrects the posture so that the image of the falling glass 25 is picked up from the best position and posture.

  In step 32, processing for initializing the data storage area and the number of data is performed together only for the first time.

  4A and 4B show images of the falling glass 25 in which the processing range is set.

  The image for which the processing range is set in step 32 in FIG. 2 is then smoothed (step 33).

  The smoothing of the image in this step 33 is performed to smooth the image, and is a pre-process for detecting the position of the figure (meaning the falling glass 25) in the image in a later step.

  As a result of the smoothing, in a noise-containing image, a dark background area z becomes uniform with bright spots and the like that are present due to noise, and the entire area z becomes a continuous flat area with little unevenness. . Similarly, in the bright area x where the glass is imaged, dark spots and the like that exist due to noise are uniform with the surrounding bright area, and the entire area x is a continuous planar area with little unevenness.

  There are various methods for smoothing processing. For example, a method of dividing an image into regions where values can be regarded as almost constant by region division processing, or a place where the luminance value of an image changes suddenly by edge detection processing is detected. A method or the like may be used.

  The image smoothed for each divided mesh region is binarized by a known discriminant analysis method (step 34).

  In the captured image, the portion where the falling glass 25 is reflected is a bright region x due to the brightness of the molten glass shown in FIGS. 4 (a) and 4 (b). The background portion is a dark area z. The binarization in step 34 is to binarize the two regions x and z into a dark region Z that is lower than the threshold and a bright region X that exceeds the threshold.

  The binarized image data of the falling glass 25 is subjected to contraction (step 35) and expansion (step 36) processing for removing image noise and quantization noise.

  Shrinkage and expansion processing is a filtering process for clarifying the unclear boundary between the binarized image background and the falling glass 25 and facilitating the left / right edge detection processing of the falling glass in step 37 described later. is there.

  This contraction processing is a processing technique in which all the pixels in the boundary portion of the figure obtained by binarization (that is, the region X which is the image of the falling glass 25) are deleted, and the boundary portion of the image is removed by one skin.

  In contrast to the contraction process, the expansion process adds one pixel in the area Z direction from the pixel of the boundary part of the obtained figure (that is, the area X that is the image of the falling glass 25), and the boundary part of the image is divided into one skin. This is a fattening method.

  By performing the contraction and expansion processing, most of the noise components in the binarized image are removed, and there is an effect that the left edge portion and the right edge portion of the glass, which is the boundary portion between the image background and the falling glass, are sharpened. .

  The data processed in this way is stored in the data storage unit 26 and processed in the next steps 37 and 38.

  The data processing unit 11 that has received the processing result of the image processing board 13 performs left / right edge detection processing (step 37).

  The data processing unit 11 detects the left edge portion and the right edge portion of the falling glass 25 from the processing result input from the image processing board 13.

  A procedure for detecting the left edge portion and the right edge portion will be described with reference to FIG. FIG. 4A shows a normal image of the falling glass 25 in which the processing range is set. FIG. 4B shows an image at the time of abnormality of the falling glass 25 in which the processing range is set (the falling glass 25 is shaken).

  An example of the detection procedure will be described by paying attention to the data of the horizontal scanning line SC1 in FIG.

First, when the number of data is larger than 0 and the previous detection of the left edge portion is normally performed, the search start point of the left side candidate position is set to the previous left edge portion −50 pixels (D ₀ in the figure). Designated distance). However, it is a precondition that the search start point is 0 or more. In other cases, the left search start point is set to zero.

On the horizontal scanning line SC1, the position where the luminance of the pixel changes and becomes brighter from the left side search start point D ₀ (or 0) toward the right end of the image with respect to the pixel which is a binarized image from the left to the right. (In this case, this is the scanning direction). That is, when a point that becomes the bright region X from the dark region Z is detected, this point is set as the left side candidate position a of the falling glass 25.

  If no detection is made even after searching to the right end of the screen, the previous detection position is set as “failure” and the process ends.

Next, when the number of data is larger than 0 and the previous right edge portion has been successfully detected, the search start point of the right side candidate position is the previous right edge portion + 50 pixels (predetermined by E ₀ in the figure). Distance). However, the search start point is on the left side of the right end of the screen. In other cases, the search start point is the right end of the image.

From the right side search start point E ₀ (or right end) to the left side candidate position a with respect to the pixels on the scanning line SC1, the pixel brightness changes from the right side to the left side (here, this Is the direction opposite to the scanning direction). That is, when a point that becomes the bright region X from the dark region Z is detected, this point is set as the right side candidate position b of the falling glass 25.

  If no detection is made even after searching to the left side candidate position a, the previous detection position is set as “failure” and the process ends.

  Here, the scanning of the two scanning lines SC1 and SC2 having different positions is shown as an example. If the scanning is performed in the horizontal direction, it is not necessary to use only the above scanning method. A scanning method that is easy to use may be used.

  For example, the left side candidate position a and the right side candidate position b may be detected by scanning the scanning lines SC1 and SC2 horizontally from left to right.

The distance between the detected left side candidate position a and the right side candidate position b is calculated, and if it is within the threshold width_th, the left side candidate position a is referred to as the left edge portion A ₀ and the right side candidate position b is set to the right It referred to the edge portion B _0.

  In other cases, the previous detection position is set as “failure” and the process ends.

Next, the width of the falling glass 25 is calculated from the difference between the detected left edge portion A ₀ and right edge portion B ₀ . When the calculated width of the falling glass 25 is larger than the threshold width_th, it is determined as “failure” to acquire the flowing width of the falling glass 25.

When the difference between the left edge portion A ₀ and the right edge portion B ₀ is smaller than the threshold width_th, it is determined that the flow width of the falling glass 25 is “successful”.

  Here, when the flow width changes, the flow width of the flowing glass 25 actually changes due to the change in the heating temperature of the melting furnace 21, or the camera 3 facing the flowing glass 25 flows down. In some cases, the glass 25 may be perceived as a phenomenon in which the flow width of the glass changes due to the glass 25 swaying in the front-rear direction of the camera 3.

  If it is determined that the width acquisition of the falling glass 25 is “successful”, the data processing unit 11 next performs a process of calculating the center position of the falling glass 25 (step 38).

The addition average of the position coordinates of the left edge part A ₀ and the right edge part B ₀ detected earlier is taken, and the calculated value is set as the coordinate C ₀ of the center position of the falling glass 25.

For the other scanning line SC2, the left edge portion A ₁ and the right edge portion B ₁ of the falling glass 25 are detected and the center position coordinate C ₁ is calculated in the same manner as described above.

  The detected and calculated result is stored in the data storage unit 16, and 1 is added to the counter of the number of data. The previous position detection is “success”.

  Next, the calculation procedure of the number of times of shaking of the falling glass 25 in the data processing unit 11 will be described.

  Since there are various types of shaking of the falling glass 25, various types of shaking detection can be considered. Here, three typical types of shaking detection methods will be described.

Shake detection method (1)
The center position of the flow nozzle 66 in the lower part 21a of the melting furnace shown in FIG. 7 is indicated by the nozzle center N in FIGS. 4 (a) and 4 (b).

The difference in the horizontal axis direction between the nozzle center N and C ₁ (or C ₀ ), which is a predetermined value, is obtained, which means a deviation of the falling glass 25 from the falling nozzle 66.

  The obtained difference is compared with a preset threshold value move_th1, and if the difference is equal to or greater than the threshold value move_th1, it is determined that the falling glass 25 has been shaken. Add 1 to the number of swings (1).

Shake detection method (2)
The difference in the horizontal axis direction between the upper and lower center positions C ₀ and C ₁ of the falling glass 25 shown in FIG. This means that the position of the stream of falling glass 25 is displaced up and down from the center position C _0.

  The obtained difference is compared with a preset threshold move_th2, and if the difference is equal to or greater than the threshold move_th2, it is determined that the falling glass 25 has been shaken. Add 1 to the number of shaking (2).

Shake detection method (3)
First, the shaking of the falling glass 25 is initialized (implemented only for the first time).

  After the initialization, the latest center position is taken out from the data storage unit 16.

The latest center position data is, for example, the coordinates C ₂ of the center position obtained in FIG.

  Next, the average value κ of the center position of the falling glass 25 for 100 times calculated in the past is acquired from the data storage unit 16. In the case where the average value κ of the center position of the falling glass 25 for the past 100 times or the data for the past 100 times is not stored in the data storage unit 16, all the data of the center position stored in the data storage unit 16 is extracted. The average value κo of all the center position data is calculated.

The latest center position C ₂ is compared with the average value κ or average value κo of the past center position, and if the difference is equal to or greater than a preset swing threshold move_th3, the falling glass 25 is shaken Judge. Add 1 to the number of shakings (3).

FIG. 4A shows a state in which there is no shaking, and if this state continues 100 times continuously, the average value κ is equal to the position of C ₀ (or C ₁ ) in the figure.

When the shaking as shown in FIG. 4B occurs in the falling glass 25, it can be seen that the difference between the average value κ and the position of C ₂ is the shaking of the falling glass 25.

  An example of counting the number of times of shaking (3) of the falling glass 25 is shown in FIG.

  In the figure, the horizontal axis represents the elapsed time (unit: minutes) at which the falling glass 25 is injected into the canister 22, and the vertical axis represents the number of times (unit: times) the shaking of the falling glass 25.

  The figure is a characteristic diagram when an abnormality has occurred in the melting furnace 21, and it can be seen that a large amount of shaking occurs particularly around the elapsed time of 150 minutes.

  The frequent occurrence of the shaking of the falling glass 25 is that the glass heating in the melting furnace 21 is not normally performed, the viscosity of the falling glass 25 becomes non-uniform due to abnormal heating, etc., and the flow rate, flow rate, etc. of the falling glass 25 are not uniform. It shows that it is stable.

  When the viscosity of the falling glass 25 is increased due to insufficient heating, a phenomenon in which a part of the falling glass 25 swells as shown by a one-dot chain line in FIG.

  Such a tendency of shaking is information that estimates the cause of shaking by separately counting and storing the shaking of the falling glass 25 detected by the shaking detection methods (1), (2), and (3). It can be.

In addition to the center positions C ₀ , C ₁ , C ₂ , the data of the left edge portions A ₀ , A ₁ , A ₂ , right edge portions B ₀ , B ₁ , B _2, etc. are stored and calculated in this way. Abnormalities such as bulging of the falling glass, generation of a glass lump, and twisting may be detected.

  Based on such data obtained in the data processing unit 11, an abnormality of the falling glass 25 can be detected. This is whether or not some abnormality is occurring in the melting furnace lower part 21a shown in FIGS. This is a sign that signs are appearing.

  Further, in the detection of the left edge portion and the right edge portion in step 37, even if failure of edge detection occurs frequently or continues, there is some abnormality in the flow down of the flow down glass 25 (for example, the flow down nozzle shown in FIG. 7). 66 may be clogged or broken (abnormality due to breakage), and this may be counted and used as abnormality determination information.

  By accumulating the obtained data and observing the trend of these data, it is possible to estimate the failure location (for example, the main electrode 61, the bottom electrode 62, the furnace bottom auxiliary electrode 63, the flow-down nozzle 66, etc.) or an abnormality. It is also possible to grasp the signs.

  The calculation processing results of the image processing board 13 and the data processing unit 11 are displayed on the display 17 and printed on the printer 18 as necessary.

  When there is a request from the remote operation system 6 via the cable 7, the calculation results of the image processing board 13 and the data processing unit 11 are transmitted to the remote operation system 6 via the cable 7, and the remote operation system Or displayed on a display provided in the remote control system.

  In this way, the center position of the falling glass 25 and the presence / absence of shaking are determined, and the process ends (step 39).

  In the falling glass abnormality monitoring device 1, by detecting the state of the falling glass 25 from the image, it is possible to detect the occurrence of shaking of the falling glass 25 as an abnormality. When the number of shakes of the falling glass 25 increases, an alarm for occurrence of an abnormality can be issued based on the detected data, or a sign of the occurrence of an abnormality can be quickly detected and displayed on a display or the like.

  As described above, by using the falling glass anomaly monitoring device 1 to monitor the state of the falling glass 25, which has relied on the mastery of the monitoring operator, the state of the falling glass 25 can be determined quantitatively by determining coordinates. It is now possible to quantitatively judge the state of flow and abnormalities from image data.

  Further, the data of the falling glass abnormality monitoring device 1 can be recorded and stored, and the accumulated data can be displayed in a graph. For this reason, by comparing with past normal state data and abnormal state data, it is possible to accurately grasp the precursor of the flow glass 25 abnormality in the melting furnace 21, or to perform abnormality prediction and abnormality determination without variation. Is possible.

  The embodiment of the present invention has industrial applicability that it can be widely applied not only to monitoring abnormalities of falling glass but also to monitoring various facilities that melt and flow down.

It is explanatory drawing which shows the falling glass abnormality monitoring apparatus which is this Embodiment, and its installation state. It is a flowchart which shows the image processing procedure of the falling glass abnormality monitoring apparatus which is this Embodiment. It is a state figure which shows the image which image | photographed the falling glass with the camera. FIG. 4A is a schematic view showing the state of the falling glass. FIG. 4B is a schematic diagram showing a state of shaking of the falling glass. It is a characteristic view which shows the frequency | count of shaking of flowing-down glass. It is explanatory drawing which shows the installation which flows down glass. It is a cross-sectional view which shows the cross section of a melting furnace lower part.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Flowing glass abnormality monitoring apparatus 2 Data processing apparatus 3 Camera 4 Cable 5 Camera control apparatus 6 Remote operation system 7, 8 Cable 11 Data processing part 12 Control signal input board 13 Image processing board 14 Image analysis software 15 Uninterruptible power supply part 16 Data Storage unit 17 Display 18 Printer 21 Melting furnace 21a Melting furnace lower part 22 Canister 23 Carriage 25 Flowing glass 30 Floor floor

Claims (7)

The molten glass obtained by mixing radioactive liquid waste in the melting furnace, allowed to flow down as a falling glass from falling nozzle, at the time of filling the falling glass canister, a stream for detecting an abnormality of the falling glass is flowing down from the melting furnace In a glass anomaly monitoring device, a camera that captures a falling glass at regular intervals and a falling glass image that is sequentially captured at regular intervals are binarized, and the binarized image is scanned horizontally. The left edge portion and the right edge portion of the falling glass are detected, the center position is calculated from the detected left edge portion and right edge portion, and the change of the center position calculated for each sequentially captured image is detected . A falling glass abnormality monitoring device comprising a data processing device for detecting shaking. The left edge portion is detected by scanning in the horizontal scanning direction, and the left edge portion is detected, and the right edge portion is detected by scanning in the opposite direction to the horizontal scanning direction. falling glass abnormality monitoring apparatus according to claim 1, wherein the detection is. The detection of the left edge part starts scanning in the scanning direction from a position a predetermined distance away from the left edge part detected last time, and the detection of the right edge part is predetermined from the right edge part detected last time to the right. The falling glass anomaly monitoring apparatus according to claim 1 or 2 , wherein scanning is started from a position apart from the distance in the direction opposite to the scanning direction. Detection of the shaking, of claims 1 to 3, wherein any difference in the horizontal direction is detected by exceeding the threshold value of the center position calculated from the center position and the left edge portion and the right edge portion of the flow-down nozzle Flowing glass abnormality monitoring device. In the detection of the shaking, the left edge portion and the right edge portion are detected by two scanning lines having different vertical positions by the scanning, and the detected left edge portion and the respective right edge are detected. The falling glass abnormality monitoring device according to any one of claims 1 to 3 , wherein a center position is calculated from a portion, and the shaking is detected when a difference in a horizontal direction between the calculated upper and lower center positions exceeds a threshold value. The falling glass abnormality monitoring device according to any one of claims 1 to 3 , wherein the shake is detected when a horizontal difference between a past center position and a latest center position exceeds a threshold value. The falling glass abnormality monitoring device according to any one of claims 1 to 6 , wherein the data processing device counts and stores the number of times of the detected shaking.

Images