JP2015092015A - Device and method for blast furnace abnormality detection - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an apparatus and a method for blast furnace abnormality detection which can detect appropriately an abnormality in the tuyere part, e.g. springing out of slag.SOLUTION: A method of detecting a blast furnace abnormality comprises arranging a camera 11 in the vicinity of the tuyere 2 of a blast furnace 1 to image a raceway part through an in-furnace monitor window 6 provided in the tuyere 2, extracting boundaries (edge) between bright parts and dark parts of the image imaged by the camera 11 and detecting a change amount (lowermost edge position change amount ΔPbtm) per specified time at the most-edge position (lowermost edge position Pbtm) of the edge. When the lowermost edge position change amount ΔPbtm is equal to or greater than a criterion threshold value TH, occurrence of springing out of slag is determined.

Description

  The present invention relates to a blast furnace abnormality detection apparatus and a blast furnace abnormality detection method for detecting an abnormality in a blast furnace tuyere.

As an operation monitoring method in the conventional blast furnace operation, for example, there is a technique described in Patent Document 1. This technology arranges measurement data (temperature, pressure, flow rate, etc.) of multiple sensors installed on the blast furnace equipment on a three-dimensional solid surface that reflects the installation position of each sensor, and spatial distribution of measurement data. The operating condition of the blast furnace is monitored based on the state and time change.
As a method for directly monitoring the inside of the furnace, for example, a method of attaching a camera to a tuyere viewing window as in the technique described in Patent Document 2, or a tuyere peeping as in the technique described in Patent Document 3, for example. There is a method to attach a fiberscope to the window.

Japanese Patent No. 4150322 Japanese Patent Application Laid-Open No. 2004-183956 JP 60-125307 A

By the way, at the time of blast furnace operation, a phenomenon that slag rises to the tip of the tuyere (hereinafter referred to as a noro spring) may occur. If such springs are generated, the tuyere may be damaged and serious trouble may be developed.
However, the techniques described in Patent Documents 1 to 3 cannot automatically detect the spring. In other words, the monitoring of the spring is performed by the operator's visual observation. For this reason, in a blast furnace having a large number of tuyere, there is a risk that the discovery of Noro spring will be delayed or overlooked.
Then, this invention makes it a subject to provide the blast furnace abnormality detection apparatus and blast furnace abnormality detection method which can detect appropriately the abnormality in a tuyere part, such as a noro spring.

  In order to solve the above problems, one aspect of a blast furnace abnormality detection device according to the present invention is a blast furnace abnormality detection device that detects an abnormality in a tuyere portion of a blast furnace, and through a monitoring window provided in the tuyere portion. An imaging unit that images the raceway unit, a bright area in which the luminance value of each pixel in the captured image captured by the imaging unit is higher than a preset luminance threshold, and the luminance value of each pixel in the captured image An edge extraction unit that extracts an edge part that is a boundary with a dark part region that is equal to or less than the luminance threshold value, and an extreme position on one side of the edge part that is extracted by the edge extraction unit in a direction opposite to the one direction When the change amount of the extreme end position detected by the extreme end position change amount detection unit and the extreme end position change amount detected by the extreme end position change amount detection unit is equal to or greater than a predetermined determination threshold, the abnormality is detected. An anomaly detector that determines that the It is characterized in that it comprises.

  Thereby, the phenomenon in which the tip of the tuyere is covered from a certain direction can be automatically detected. Examples of such a phenomenon include a noro spring in which slag rises to the tip of the tuyere and sticking of unmelted ore to the top of the tuyere. Thus, the abnormality in the tuyere can be detected appropriately. At this time, since the rate of change of the extreme end position of the edge extracted from the captured image of the tuyere is monitored, an abnormality can be detected relatively easily and, for example, the extreme edge of the edge due to a gradual temperature change A change in position can be prevented from being detected as abnormal.

Further, in the above, the extreme end position change amount detection unit is configured such that the extreme end position on the side corresponding to the downward direction of the blast furnace in the edge portion per predetermined time in the direction corresponding to the upward direction of the blast furnace. It is preferable to detect the amount of change.
Thus, it can be detected that the slag has risen to the tip of the tuyere when the lowest position of the edge extracted from the captured image has risen. Accordingly, it is possible to appropriately detect a noro spring and prevent troubles such as tuyere damage.

  Further, in the above, based on the amount of change per predetermined time of the area of the bright area, when the extreme end position in all directions of the edge portion is uniformly changed toward the center of the bright area, An extreme end position change amount correction unit that corrects the change amount of the extreme end position detected by the extreme end position change amount detection unit so as to cancel the change in the extreme end position on the one-direction side of the edge portion; Is preferred.

  As described above, when the extreme end position of an edge in only one direction changes toward the bright part region center side, the extreme end position of the edge in all directions including the one direction changes equally toward the bright part region center side. By utilizing the fact that the amount of change in the area of the bright region differs depending on the case, the amount of change in the extreme end position is corrected. Therefore, even when the extreme end positions of the edges in all directions are changed evenly due to the change in the combustion state, this can be prevented from being detected as an abnormality in the tuyere.

Further, in the above, the imaging unit is provided in each of the plurality of tuyere provided in the blast furnace, and the abnormality detecting unit is provided in at least two of the tuyere continuously provided in the plurality of tuyere. When it is determined by mouth that the abnormality has occurred within a certain period of time, it is preferable to determine that an abnormality due to noro-nosake has occurred in the blast furnace.
In this way, by using the detection results at a plurality of tuyere provided continuously, even when the field of view is narrow, it is possible to appropriately detect the noro-spring by suppressing over-detection due to changes in the combustion state, etc. it can.

Further, one aspect of the blast furnace abnormality detection method according to the present invention is a blast furnace abnormality detection method for detecting an abnormality in a tuyere portion of a blast furnace, and images a raceway portion through a monitoring window provided in the tuyere portion. , An edge serving as a boundary between a bright area where the luminance value of each pixel in the captured image is higher than a preset luminance threshold and a dark area where the luminance value of each pixel in the captured image is equal to or less than the luminance threshold When the amount of change per predetermined time in the direction opposite to the one direction of the one-end side of the part is detected and the amount of change is equal to or greater than a predetermined determination threshold, the abnormality has occurred. It is characterized by judging.
In this way, since the rate of change of the extreme edge position of the edge extracted from the captured image of the tuyere is monitored, it is possible to relatively easily detect abnormalities such as a noro spring where the slag rises to the tip of the tuyere. .

  According to the present invention, since the edge of the bright area and the dark area of the blast furnace image taken from the tuyere monitoring window is detected and the change in the lower part of the edge is monitored, abnormalities such as noro-swell are automatically detected. be able to. Further, it is possible to easily detect an abnormality as compared with the case where the operator visually monitors. Therefore, it is possible to avoid the occurrence of a serious trouble due to the abnormality of the tuyere, and the effect can be obtained in terms of safety and equipment repair cost.

1 is an overall view of a blast furnace to which a blast furnace abnormality detection device of an embodiment is applied. It is a figure which shows the installation position of a camera. It is a figure which shows the example of the image imaged with the camera. It is a figure explaining noro spring detection logic. It is a flowchart which shows the abnormality detection process procedure in 1st Embodiment. It is a figure which shows the lowest end position variation | change_quantity (DELTA) Pbtm of the time including a noro-spring phenomenon. It is a change of the edge lower part by the change of a combustion state. It is a figure explaining the correction method of bottom end position change amount (DELTA) Pbtm. It is a flowchart which shows the abnormality detection process procedure in 2nd Embodiment. It is a figure which shows the lowest end position variation | change_quantity (DELTA) Pbtm of the time including a combustion state change. It is a figure which shows the lowest end position variation | change_quantity (DELTA) Pbtm of the time including a noro-spring phenomenon. It is a block diagram which shows the blast furnace abnormal apparatus in 3rd Embodiment. It is a flowchart which shows the abnormality detection process procedure in 3rd Embodiment. It is a figure which shows the difference of the captured image by the change of a combustion state in the state where the visual field of the monitoring window in a furnace became narrow. It is a figure which shows the lowest end position variation | change_quantity (DELTA) Pbtm of the time including the noro-swell phenomenon in two tuyere.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
(First embodiment)
FIG. 1 is an overall view of a blast furnace to which the blast furnace abnormality detection device of the present embodiment is applied.
As shown in FIG. 1, a blower pipe (blow pipe) 3 for blowing hot air from a hot stove is connected to the inside of the tuyere 2 of the blast furnace 1 and penetrates through the blower pipe 3. The lance 4 is installed. From the lance 4, fuel such as pulverized coal, oxygen and city gas is blown into the furnace.
A combustion space called a raceway 5 exists in the coke deposit layer in front of the tuyere 2 in the direction of blowing hot air, and coke combustion and gasification (reduction of iron ore, that is, ironmaking) are mainly performed in this combustion space. .

Further, as shown in FIG. 2, an in-furnace monitoring window 6 is formed at the tuyere for the operator to monitor the inside of the furnace. A camera 11 for imaging the raceway 5 through the in-furnace monitoring window 6 is installed in the vicinity of the in-furnace monitoring window 6.
FIG. 3 is a diagram illustrating an example of an image captured by the camera 11. As shown in FIG. 3, the captured image shows the silhouette of the raceway 5 and the lance 4 inside the circular shape corresponding to the tip opening of the small tuyere 2 a constituting the tuyere 2.
A captured image of the raceway section captured by the camera 11 is input to the abnormality detection section 12. The abnormality detection unit 12 detects an abnormality in which the slag rises up to the tip of the tuyere 2 using the captured image captured by the camera 11.

  Noro springs are generated when slag accumulates and the pressure in the furnace decreases. In the case where a spring is generated, the lower part of the imaging range of the camera 11 is hidden by the rising slag. Since the temperature of the slag is relatively low, the brightness of the part corresponding to the slag in the captured image decreases, and the lowest end position of the bright part in the captured image (the extreme end position on the side corresponding to the downward direction of the blast furnace) is normal As a result, the phenomenon rises.

FIG. 4 is a diagram for explaining the noro spring detection logic.
The picked-up image picked up by the camera 11 at the normal time when no spring is generated is as shown in the upper part of FIG. In addition, a boundary (edge) between a bright area where the luminance value of each pixel in the captured image is higher than a preset luminance threshold and a dark area where the luminance value of each pixel in the captured image is equal to or lower than the luminance threshold. Is extracted, the edge image is as shown in the lower part of FIG. At this time, the lowest end position Pbtm of the edge is Pbtm1.
Here, the lowest edge position Pbtm of the edge indicates the position of the lowest edge of the edge in terms of the number of pixels when the lowermost part of the captured image is 0 [pixel]. That is, the lowermost position Pbtm of the edge becomes larger as the position of the lowermost edge of the edge is higher in the captured image.

Then, when nodding occurs from this normal state, the camera 11 captures a captured image as shown in the upper part of FIG. When the edge in this captured image is extracted, it becomes as shown in the lower part of FIG. Thus, the lower part of the captured image is darkened by the slag rising to the tip of the tuyere 2, and the lowest end position Pbtm of the edge is increased by the amount of the increased slag. If the lowest end position Pbtm of the edge at this time is Pbrm2, the amount of change is ΔPbtm (= Pbtm2−Pbtm1).
Therefore, when the lowest end position change amount ΔPbtm is equal to or greater than the preset determination threshold value TH, it can be determined that the springing is occurring.

Therefore, the abnormality detection unit 12 detects the edge of the image inside the tuyere and detects the springing by monitoring the change in the lower part of the edge. The detection result by the abnormality detection unit 12 is displayed on the monitor 13 and notified to the operator. In addition, the alarm device 14 is activated when the abnormality detecting unit 12 detects the nodding, and this is notified to the operator.
FIG. 5 is a flowchart showing an abnormality detection processing procedure executed by the abnormality detection unit 12. This abnormality detection process is repeatedly executed every predetermined time (for example, every 0.3 seconds). First, in step S <b> 1, the abnormality detection unit 12 acquires a captured image captured by the camera 11.

Next, in step S2, the abnormality detection unit 12 performs a binarization process on the captured image acquired in step S1 using a preset brightness threshold for edge detection, and converts the image data that exceeds the brightness threshold. A boundary (edge) between the corresponding bright part and the dark part corresponding to the image data equal to or lower than the luminance threshold is extracted.
Next, in step S3, the abnormality detection unit 12 detects the lowermost edge position Pbtm of the edge based on the edge extraction result in step S2, and proceeds to step S4.
In step S4, the abnormality detector 12 detects the lowest edge position Pbtm (n) of the edge in the current sampling process detected in step S3 and the lowest edge position Pbtm (n−1) of the edge detected in the previous sampling process. ) And the lowest end position change amount ΔPbtm (= Pbtm (n) −Pbtm (n−1)) is calculated.

Next, in step S5, the abnormality detection unit 12 determines whether or not the lowest end position change amount ΔPbtm calculated in step S4 is equal to or greater than a determination threshold value TH. Here, the determination threshold TH is a positive value. If ΔPbtm ≧ TH, the process proceeds to step S6, and if ΔPbtm <TH, the process proceeds to step S7 described later.
In step S6, the abnormality detection unit 12 determines that no-swell has occurred (abnormality detection), displays the abnormality detection result on the monitor 13, and activates the alarm device 14 before terminating the abnormality detection processing. .
In step S7, the abnormality detection unit 12 determines that no spring has occurred (abnormality non-detection), and ends the abnormality detection process.

Hereinafter, the abnormality detection process in the tuyere will be described using a specific example.
First, the abnormality detection unit 12 first acquires a captured image of the raceway unit captured by the camera 11 installed in the specific tuyere 2 (step S1 in FIG. 5), and then the edge portion in the acquired captured image. Is extracted (step S2). Subsequently, the lowest end position Pbtm of the extracted edge is detected (step S3), and the change amount ΔPbtm of the lowest end position Pbtm is calculated (step S4). As a result of actually detecting the lowest end position Pbtm and the lowest end position change amount ΔPbtm, the result shown in FIG. 6 was obtained.

FIG. 6 is a diagram showing time-series data of the lowermost end position Pbtm and the lowermost end position change amount ΔPbtm at a time including the noro phenomenon. In FIG. 6, the symbol a indicates the lowermost position Pbtm, and the symbol b indicates the lowermost position change amount ΔPbtm.
As shown in the portions surrounded by the alternate long and short dash lines A and B in FIG. 6, the bottom end position Pbtm rises abruptly compared with the normal time at the time when the noro spring occurs, and as a result, the bottom end position change amount ΔPbtm. Becomes equal to or greater than the determination threshold TH indicated by a broken line (Yes in step S5).
Therefore, by performing threshold processing using the determination threshold TH on the lowest end position change amount ΔPbtm, it is possible to detect the springing easily and appropriately.

As described above, in this embodiment, the camera 11 captures an image of the raceway portion, and monitors the amount of change below the edge, which is the boundary between the bright and dark portions in the captured image. In other words, when noro springs occur, the lowermost position of the edge is taken advantage of the fact that the lower part of the blast furnace image taken from the in-furnace monitoring window 6 (the side corresponding to the lower direction of the blast furnace) is hidden by the nose and darkens. A noro spring is detected with an increase in Pbtm. As a result, it is possible to automatically detect the noro spring.
At this time, the threshold value process is performed on the change amount ΔPbtm per predetermined time of the lowest end position Pbtm of the edge to detect the springing. Accordingly, it is possible to appropriately detect a phenomenon in which the lowermost edge position Pbtm of the edge suddenly rises, and it is possible to suppress erroneous detection of a moderate temperature change or the like of the raceway portion as a spring.

In addition, when noro spring is detected, this is notified to the operator, so the operator may take action to adjust the operating conditions such as increasing the amount of hot air blown to increase the pressure in the furnace. it can.
In this way, it is possible to automatically detect noro springs and appropriately handle abnormal conditions, preventing serious troubles such as tuyere damage, which is effective in terms of safety and equipment repair costs. It is done.
In the above, the camera 11 corresponds to the imaging unit, step S2 in FIG. 5 corresponds to the edge extraction unit, steps S3 and S4 correspond to the extreme end position change amount detection unit, and steps S5 and S6 detect abnormality. Corresponds to the department.

(Second Embodiment)
Next, a second embodiment of the present invention will be described.
In the second embodiment, correction is applied to the lowermost position change amount ΔPbtm so as not to detect the change in the lowermost position Pbtm of the edge due to the change in the combustion state when detecting the noro-swell. .
At the time of operation, when the combustion state in the furnace changes suddenly due to the change in the blowing state from the lance 4, the brightness in the image captured by the camera 11 changes, and the shape of the edge portion extracted from the captured image changes. For example, if the captured image changes from the state shown in the upper part of FIG. 7A to the state shown in the upper part of FIG. 7B due to the change in the combustion state, the edge portion extracted from the captured image is shown in FIG. The state shown in the lower part of a) changes to the state shown in the lower part of FIG.

That is, the extreme end position of the edge portion changes substantially evenly in the vertical and horizontal directions. In this way, even if the combustion state changes, the phenomenon that the bottom end position of the edge rises occurs.Therefore, if only the amount of change in the bottom end position of the edge is monitored, the change in the combustion state is I will judge.
Therefore, in the present embodiment, the lowermost position change amount ΔPbtm is corrected so as to cancel out the change in the lower edge due to such a change in the combustion state, and the determination threshold TH is applied to the corrected lowermost position change amount ΔPbtm. Threshold processing using is performed.

  In the case of a noro spring, only the lower part of the captured image is dark because it is hidden by the noro, whereas in the case of a change in the combustion state, all of the upper, lower, left and right are darkened almost uniformly. That is, in the case of noro spring, the area change amount of the bright part is smaller than that in the case of the combustion state change. Therefore, by using the difference in the area change amount of the bright part and correcting the bottom end position change amount ΔPbtm using the area change amount of the bright part, it is possible to discriminate the difference between the normal spring and the change in the combustion state. To.

FIG. 8 is a flowchart illustrating an abnormality detection processing procedure executed by the abnormality detection unit 12 according to the second embodiment. This abnormality detection process is the same as the abnormality detection process of FIG. 5 except that the process of step S11 is added after step S4 of FIG. Therefore, here, the description will focus on the different parts of the processing.
In step S11, the abnormality detection unit 12 corrects the lowest end position change amount ΔPbtm calculated in step S4, and proceeds to step S5. In this step S11, first, a correction value C = (√A−√B) / √π for correcting the lowest end position change amount ΔPbtm is calculated. Here, A is the area of the bright part in the captured image acquired by the previous sampling process, and B is the area of the bright part in the captured image acquired by the current sampling process.
Next, the abnormality detection unit 12 subtracts the correction value C from the lowest end position change amount ΔPbtm calculated in Step S4, and the result is used as the corrected bottom end position change amount ΔPbtm and proceeds to Step S5.

Hereinafter, the correction method based on the bright area will be described in detail.
FIG. 9A is a diagram schematically showing changes in the bright area due to changes in the combustion state. Here, the bright part is assumed to be circular.
As described above, the change in the extreme end position of the edge due to the change in the combustion state occurs substantially uniformly in the vertical and horizontal directions. If the bright part radius is changed from R to r due to the change of the combustion state, the difference of the bright part area is (A−B) = π (R2−r2), the difference of the lowest end position (the change amount of the lowest end position) ΔPbtm = (R−r).
Therefore, the lowermost position change amount ΔPbtm is expressed as follows using the bright area.
ΔPbtm = R−r = √π (R−r) / √π = √π (√R2−√r2) / √π
= (√ (πR2) −√ (πr2)) / √π
= (√A−√B) / √π (1)
Therefore, if the bright part radius is halved from 2r to r due to a change in the combustion state, the lowermost position change amount ΔPbtm = r is a value (√A−√B) expressed using the bright part area. ) / √π.

Next, consider a case where only the lowermost position Pbtm of the bright part has risen due to springing. FIG.9 (b) is a figure which shows typically the change of the bright part area by Noro spring. As shown in FIG. 9B, when the lowest end position Pbtm of the circular bright part having a radius of 2r is increased by r, the bright part area A = 4πr2 before the occurrence of the noro-swell, the bright part when the noro-swell occurs. The area B = (8πr2) / 3 + √3r2.
Therefore, the bright area difference (√A−√B) can be expressed as follows.
√A−√B = √ (4πr2) −√ ((8πr2) / 3 + √3r2)
<√ (4πr2) −√ ((8πr2) / 3 = r√π (2-2√6 / 3)
……… (2)
Here, since (2-2√6 / 3) <1, the above equation (2) can be expressed as follows.
√A−√B <r√π (3)
Therefore, when the lowest end position Pbtm rises by r due to springing, the lowest end position change amount ΔPbtm = r is greater than the value (√A−√B) / √π expressed using the bright area. Become.

That is, when the extreme end position of the edge in all directions including the downward direction changes uniformly toward the center of the bright portion due to the change in the combustion state, the change in the lowest end position ΔPbtm is (√A−√B) / √ The result of subtracting π is almost zero. On the other hand, when only the lowermost position of the edge is changed (increased) due to springing, it does not become 0 even if (√A−√B) / √π is subtracted from the lowermost position change amount ΔPbtm.
Therefore, (√A−√B) / √π is used as the correction value C of the lowermost position change amount ΔPbtm, and the correction value C is subtracted from the lowermost position change amount ΔPbtm calculated based on the captured image of the camera 11. Thus, the lowermost position change amount ΔPbtm is corrected. Thereby, the change in the lowest end position Pbtm caused by the change in the combustion state can be canceled out.

FIG. 10 is a diagram illustrating the lowest end position change amount ΔPbtm at the time including the combustion state change. In the figure, the broken line is the bottom end position change amount ΔPbtm before correction, and the solid line is the bottom end position change amount ΔPbtm after correction using the change amount of the bright area.
When the combustion state changes, the bright area in the captured image of the camera 11 changes, and when the lowermost edge position Pbtm of the edge rises, the lowermost edge that is greater than or equal to the determination threshold TH as shown by the portion surrounded by the alternate long and short dash line C A position change amount ΔPbtm is detected. Therefore, if the detected lowermost position change amount ΔPbtm is compared with the determination threshold value TH as it is, it is erroneously determined that the spring has occurred.

  On the other hand, in the present embodiment, the detected bottom end position change amount ΔPbtm is corrected by a correction value C = (√A−√B) / √π using the change amount of the bright area in the captured image. As a result, it is possible to obtain the bottom end position change amount ΔPbtm that cancels out the change in the bottom end position due to the combustion state change. As a result of the correction, the lowest end position change amount ΔPbtm after the correction is lower than the determination threshold value TH as shown by the portion surrounded by the alternate long and short dash line C. For this reason, even when compared with the determination threshold value TH, noro-no-suga is not detected. In this way, it is possible to prevent a change in the combustion state from being erroneously determined as a spring.

FIG. 11 is a diagram illustrating the lowest end position change amount ΔPbtm at the time including the noro-swell phenomenon. In the figure, the broken line is the bottom end position change amount ΔPbtm before correction, and the solid line is the bottom end position change amount ΔPbtm after correction using the change amount of the bright area.
As shown in FIG. 11, in the case where the spring is actually generated, the correction value C = (√A−√ using the change amount of the bright area in the captured image from the detected change amount ΔPbtm of the lowermost end. Even if B) / √π is subtracted, the bottom end position change amount ΔPbtm after correction does not change significantly. Therefore, the corrected lowermost position change amount ΔPbtm at the time point when the springing occurs is equal to or greater than the determination threshold TH as shown in the portion surrounded by the alternate long and short dash line D. Accordingly, it is possible to appropriately detect the noro spring.

As described above, the bottom end position change amount ΔPbtm is corrected based on the area change amount of the bright area in the captured image, and the change in the bottom end position due to the combustion state change is canceled out. The phenomenon that the lowest end position Pbtm rises can be regarded as non-swelling non-detection. As a result, overdetection can be suppressed.
Further, by assuming that the bright area is a circle, it is possible to set an appropriate correction value C for correcting the lowermost position change amount ΔPbtm. That is, the bright area in the captured image actually captured by the camera 11 is not a perfect circle, but an appropriate abnormality detection result can be obtained even assuming a circle.
In the above, step S11 in FIG. 8 corresponds to the extreme end position change amount correction unit.

(Third embodiment)
Next, a third embodiment of the present invention will be described. The blast furnace abnormality detection device according to the third embodiment includes two cameras 11a and 11b, an abnormality detection unit 12, a monitor 13, and an alarm device 14, as shown in FIG. The blast furnace 1 is provided with a plurality of tuyere 2 along the outer periphery of the blast furnace 1.

The cameras 11a and 11b are the same as the cameras 11 of the first and second embodiments, and are respectively provided in two tuyere 2 provided continuously along the outer periphery of the blast furnace 1.
The abnormality detection unit 12 detects an abnormality caused by the springing by a detection logic described later.
The monitor 13 and the alarm device 14 are the same as those in the first and second embodiments.

Next, the abnormality detection logic caused by the noro spring by the abnormality detection unit 12 in the present embodiment will be described.
As shown in FIG. 13, first, the abnormality detection unit 12 determines an abnormality for each tuyere 2 (S21). The abnormality in step S21 indicates whether or not the noro spring is detected, and it is determined that the abnormal is detected when the noro spring is detected as in the first and second embodiments. In step S <b> 21, the noro springs of the two tuyere 2 continuously provided are detected from the captured images of the cameras 11 a and 11 b using the blast furnace abnormality detection method according to the first and second embodiments. That is, when using the blast furnace abnormality detection method according to the first embodiment, the abnormality detection unit 12 performs the processing of Step S1 to Step S7 shown in FIG. Judgment of whether or not noro spring is detected. Moreover, when using the blast furnace abnormality detection method according to the second embodiment, the abnormality detection unit 12 performs the processes of steps S1 to S7 and S11 illustrated in FIG. Judgment is made individually on whether or not noro spring is detected. The determination of abnormality in step S21 is continuously performed at predetermined intervals (for example, every 1 second).

Next, the abnormality detection unit 12 determines whether or not an abnormality has been detected within a certain period of time from both of the two consecutive tuyere 2 from the determination result of step S21 (S22). For example, the abnormality detection unit 12 determines whether an abnormality is detected in the two tuyere 2 within 2 minutes as a certain period.
If an abnormality is detected in both of the two tuyere 2 in the determination in step S22, the abnormality detecting unit 12 determines that the noro-spring has occurred (abnormality detection), and monitors the abnormality detection result due to the noro-spring 13 And the alarm device 14 is activated (S23).

On the other hand, if no abnormality is detected in the two tuyere 2 in the determination in step S22, the abnormality detection unit 12 determines that no spring has occurred (abnormality is not detected). At this time, an abnormality non-detection result may be displayed on the monitor 13.
After steps S22 and S23, the abnormality detection process is terminated.
As described above, in the third embodiment, in order to suppress over-detection caused by various factors such as narrowing of the tuyere image field, Based on the detection result, an abnormality caused by the spring is detected.

  In recent years, the field of view of the in-furnace monitoring window 6 may be narrowed due to the presence of a pulverized coal (PC) blowing lance or the axis deviation of the blow pipe. For example, when the upper and lower visual fields become narrow as in the example shown in FIG. 14A, it is necessary to detect the noro spring from the bright portion area having high luminance below the lance 4. However, when the area of the bright area changes due to the change in the combustion state, the upper part is hidden by the lance 4 in addition to the narrow field of view as shown in FIG. Therefore, only the bottom part may change greatly. At this time, in the blast furnace abnormality detection method according to the first embodiment, overdetection is likely to occur as described in the second embodiment. Further, in the blast furnace abnormality detection method according to the second embodiment, only the lowermost part of the bright region changes greatly, so that it becomes easy to detect a change in the combustion state as a spring. For this reason, overdetection easily occurs in the blast furnace abnormality detection method according to the second embodiment. Furthermore, since the field of view changes as equipment is changed every time the wind is off, it is difficult to remove only the tuyere 2 with a narrow field of view even if the field of view of the in-furnace monitoring window 6 is narrow.

  Since Noro is a liquid, it is difficult to consider that the level of Noro rises locally, and when Noro Spring occurs, Noro Spring occurs at two tuyere 2 that are installed in a short period of time. For this reason, in this embodiment, about two tuyere 2 installed continuously, it is judged whether anomaly was detected in both tuyere 2 within a fixed period, and abnormalities caused by Noro springs are detected. To detect. Moreover, since the abnormality due to the change in the area of the bright area occurring only in one tuyere 2 is an overdetection, such an abnormality can be excluded in this embodiment. Therefore, even in the blast furnace 1 having a structure in which over-detection is likely to occur frequently, such as the field of view of the in-furnace monitoring window 6 is narrow, over-detection can be suppressed and a normal spring can be detected appropriately.

  FIG. 15 shows time-series data of the bottom end position change amount ΔPbtm at the time including the noro-swell phenomenon in two tuyere 2 provided in succession. In FIG. 15, the data indicated by c and d indicate the value of the lowest end position change amount ΔPbtm in the two tuyere 2, and the area indicated by D indicates the time when the noro-swell occurs. Further, in the example shown in FIG. 15, in step S2, the same method as in the first embodiment is used to determine whether or not the nodding is detected for each of the two tuyere 2 and to determine the maximum. The threshold value for the lower end position change amount was set to 15 pixels. Furthermore, in step S22, the period for determining whether or not an abnormality was detected in both tuyere 2 was set to 2 minutes.

  As shown in FIG. 15, at the time when the noro-spring was generated, the lowermost position change amount ΔPbtm of both c and d exceeded the threshold value within the period of 2 minutes, and the anomaly caused by the noro-spring could be detected. . Further, in the example shown in FIG. 15, since the bottom end position change amount ΔPbtm of both c and d does not exceed the threshold value within 2 minutes, the noro-spring detection (the bottom end position change in the individual tuyere 2) is detected. The amount ΔPbtm exceeds the threshold value) is excluded as overdetection. Therefore, according to the present embodiment, it was confirmed that even in the blast furnace 1 having a structure in which overdetection is likely to occur frequently, it is possible to appropriately detect the noro spring while suppressing overdetection.

(Modification)
In each of the embodiments described above, the case where the amount of change in the lowermost edge position of the edge is monitored has been described. However, the amount of change in the uppermost edge position of the edge can also be monitored. In this case, the descending amount per predetermined time of the uppermost end position of the edge is monitored, and it is determined whether or not the descending amount is equal to or greater than a preset determination threshold value. Thereby, for example, it is possible to determine whether or not a phenomenon has occurred in which unmelted ore attached to the top of the tip of the tuyere falls and attempts to close the tuyere.

  Furthermore, the present invention can be applied to monitoring the amount of change in the extreme end position on either the left or right side of the edge. Since partial clogging due to sticking of unmelted ore may occur even from the lateral direction of the tip of the tuyere, this can be detected by monitoring the amount of change in the extreme end position on either the left or right side of the edge . Normally, pulverized coal is blown from the lance 4 together with oxycol. However, if the combustion state is normal, the lance is burned at the tip of the lance and the pulverized coal cannot be seen (the lance tip is bright). However, when the combustion state deteriorates, the pulverized coal may appear black. By monitoring the amount of change in the extreme end position on either the left or right side of the edge, it can also be determined.

  Further, in the third embodiment, the abnormality detection unit 12 detects the abnormality caused by the noro-spring from the captured images of the cameras 11a and 11b provided in the two tuyere 2, but the present invention is in such an example. It is not limited. For example, the abnormality detection unit 12 may detect an abnormality caused by the spring from the captured images of the plurality of cameras 11 provided in the plurality of tuyere 2 respectively. At this time, in step S <b> 22, abnormality may be detected by determining whether or not noro-swell is detected in at least two continuous tuyere 2 out of the plurality of tuyere 2.

  DESCRIPTION OF SYMBOLS 1 ... Blast furnace, 2 ... Air blow pipe, 3 ... Tuyere, 4 ... Lance, 5 ... Raceway, 6 ... Monitoring window in furnace, 11, 11a, 11b ... Camera, 12 ... Abnormality detection part, 13 ... Monitor, 14 ... Alarm device

Claims (5)

A blast furnace abnormality detection device that detects an abnormality in the tuyere of the blast furnace,
An imaging unit that images the raceway unit through a monitoring window provided in the tuyere,
A bright area where the luminance value of each pixel in the captured image captured by the imaging section is higher than a preset luminance threshold value, and a dark area where the luminance value of each pixel in the captured image is equal to or less than the luminance threshold value An edge extraction unit for extracting an edge part serving as a boundary;
An extreme end position change amount detection unit that detects a change amount per predetermined time in a direction opposite to the one direction of the extreme end position on one direction side of the edge portion extracted by the edge extraction unit;
An abnormality detection unit that determines that the abnormality has occurred when a change amount of the extreme end position detected by the extreme end position change amount detection unit is equal to or greater than a predetermined determination threshold value. Blast furnace abnormality detection device.   The extreme end position change amount detection unit detects the change amount per predetermined time in the direction corresponding to the upward direction of the blast furnace at the extreme end position of the edge portion corresponding to the downward direction of the blast furnace. The blast furnace abnormality detection device according to claim 1.   Based on the amount of change of the area of the bright area per predetermined time, when the extreme end position of the edge portion in all directions changes uniformly toward the center of the bright area, the edge portion An end position change amount correction unit that corrects the amount of change in the end position detected by the end position change amount detection unit so as to cancel out the change in the end position on one side. Item 3. A blast furnace abnormality detection device according to item 1 or 2. The imaging unit is provided in each of the tuyere provided in the blast furnace,
The abnormality detection unit is caused by a noro spring in the blast furnace when it is determined that the abnormality has occurred within a certain period in at least two of the tuyere provided continuously among the tuyere. The blast furnace abnormality detection device according to any one of claims 1 to 3, wherein it is determined that an abnormality has occurred. A blast furnace anomaly detection method for detecting an anomaly in a tuyere of a blast furnace,
The raceway part is imaged through a monitoring window provided in the tuyere part,
An edge portion serving as a boundary between a bright area where the luminance value of each pixel in the captured image is higher than a preset luminance threshold and a dark area where the luminance value of each pixel in the captured image is equal to or less than the luminance threshold When the change amount per predetermined time in the direction opposite to the one direction of the extracted edge portion in one direction is detected, and the change amount is equal to or greater than a preset determination threshold value A method of detecting an abnormality in a blast furnace, wherein the abnormality is determined to have occurred.