Furnace wall temperature distribution measurement system, furnace wall abnormality diagnosis system, operation method of coke oven, and repair method of coke oven

Abstract

An oven-wall temperature distribution measurement system according to the present invention includes: a thermal imaging instrument that is installed at a position at which the entire oven wall surface of a carbonization chamber disposed in a coke oven on an extruder can be imaged; an image acquisition unit that acquires image data of the oven wall surface imaged by the thermal imaging instrument; a projective transformation unit that applies projective transformation to the acquired image data of the oven wall surface; and a temperature transformation unit that eliminates influences of the emissivity due to the imaging angle in the image data to which the projective transformation has been applied and transforms said image data to temperature distribution data.

Description

Technical Field

[0001] The present invention relates to a furnace wall temperature distribution measurement system, a furnace wall abnormality diagnosis system, an operation method of a coke oven, and a repair method of a coke oven.

Background Art

[0002] A coke oven for producing coke by carbonizing coal has a structure in which a regenerator is provided at the lower part, and combustion chambers and carbonization chambers partitioned by furnace wall bricks, which are refractory bricks, are alternately arranged at the upper part. The coal charged into the carbonization chamber is carbonized by heat transfer from the combustion chambers arranged on both sides of the carbonization chamber to become coke. The red-hot coke after carbonization is pushed from the side of the pusher machine by the pusher ram of the pusher machine and discharged to the opposite side. In a chamber furnace type coke oven, a series of operations such as coal charging, coking, and coke pushing are repeated in this way. At this time, the production amount of coke increases or decreases by adjusting the residence time from coal charging to coke pushing, that is, the length of the carbonization time. In order to produce coke of a desired quality, the carbonization time needs to be longer as the carbonization temperature (furnace temperature) is lower and the furnace width is wider. Therefore, the production amount of coke can be increased or decreased by setting the carbonization temperature within the constraints that guarantee the desired quality. In addition, in order to set the carbonization temperature according to the production amount, the temperature distribution in the carbonization chamber of the coke oven is controlled by adjusting the amount of gas and air supplied to the combustion chamber. As a method for measuring the temperature distribution in the carbonization chamber of a coke oven, Patent Document 1 discloses a method using a two-color radiation thermometer attached to the pusher ram of a pusher machine.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] As disclosed in Patent Document 1, when a two-color radiation thermometer is installed on the extrusion ram, when the coke is extruded from the carbonization chamber, the two-color radiation thermometer is exposed to the ultra-high temperature environment (700 to 1200 °C) in the carbonization chamber. Therefore, the deterioration of the two-color radiation thermometer is significantly accelerated, and as a result, there is a risk of causing a failure of the two-color radiation thermometer. In addition, in order to grasp the entire temperature distribution in the carbonization chamber, in the case of a two-color radiation thermometer, since it is a line measurement of the furnace wall surface, the measurement is made in accordance with the movement line of the extrusion ram, and the measurement location is limited to a specific part. If a change occurs in the furnace wall state outside this movement line, there is a possibility that this change cannot be recognized. Therefore, it is difficult to grasp the temperature distribution of the entire furnace wall surface. Further, when performing temperature measurement of the entire furnace wall surface from outside the carbonization chamber using a radiation thermometer that measures temperature from the radiation amount such as thermography, usually, since the extrusion ram is located in front of the carbonization chamber, it is difficult to take a photograph from the front. In the image obtained by photographing the furnace wall surface obliquely, the furnace wall surface disappears at a certain vanishing point geometrically. Therefore, correction of the depth is required. At this time, there is also a problem that the radiation thermometer is affected by the emissivity depending on the angle. Therefore, conventionally, the temperature management of the coke oven is often performed by measuring the temperature of the combustion chamber, and it has been difficult to constantly measure the temperature in the carbonization chamber, identify abnormal locations, and feedback them to the operation and repair actions.

[0005] The present invention has been made in view of the above problems, and an object thereof is to reduce the failure frequency of the measuring instrument, enable measurement of the temperature distribution of the entire furnace wall surface, identify abnormal locations of the furnace wall, and provide a furnace wall temperature distribution measurement system, a furnace wall abnormality diagnosis system, an operation method of a coke oven, and a repair method of a coke oven that can feedback to operation and repair actions.

Means for Solving the Problems

[0006] In order to solve the above-described problems and achieve the object, (1) The furnace wall temperature distribution measurement system according to the present invention includes a thermal imaging measuring device installed at a position capable of imaging the entire furnace wall surface of a carbonization chamber disposed in a coke oven on an extruder, an image acquisition unit that acquires image data of the furnace wall surface imaged by the thermal imaging measuring device, a projective transformation unit that performs projective transformation on the acquired image data of the furnace wall surface, and a temperature conversion unit that eliminates the influence of emissivity caused by the imaging angle on the projected image data and converts it into temperature distribution data.

[0007] (2) The furnace wall abnormality diagnosis system according to the present invention includes a thermal imaging measuring device installed at a position capable of imaging the entire furnace wall surface of a carbonization chamber disposed in a coke oven on an extruder, an image acquisition unit that acquires image data of the furnace wall surface imaged by the thermal imaging measuring device, a projective transformation unit that performs projective transformation on the acquired image data of the furnace wall surface, a temperature conversion unit that eliminates the influence of emissivity caused by the imaging angle on the projected image data and converts it into temperature distribution data, and an abnormality determination unit that determines an abnormality of the furnace wall surface based on the image data projected by the projective transformation unit and / or the temperature distribution data converted by the temperature conversion unit.

[0008] (3) The furnace wall abnormality diagnosis system according to the present invention is the invention of (2) above, wherein the abnormality determination unit determines that a portion where the temperature exceeds a predetermined first threshold value or a portion lower than a second threshold value is abnormal.

[0009] (4) The furnace wall abnormality diagnosis system according to the present invention is the invention of (2) above, wherein the abnormality determination unit registers a plurality of pieces of information on abnormal shapes and abnormal sizes in advance as templates for the image data, and determines the abnormality level from the matching ones.

[0010] (5) The furnace wall abnormality diagnosis system according to the present invention, in the invention of (2) above, the abnormality determination unit uses, as input, the image data projection-transformed by the projection transformation unit, and outputs, as output, a first learned model machine-learned with the normal state and abnormal state of the furnace wall surface, and a second learned model machine-learned with the temperature distribution data as input and the normal state and abnormal state of the furnace wall surface as output, to perform an abnormality determination.

[0011] (6) The furnace wall abnormality diagnosis system according to the present invention, in the invention of (2) above, the abnormality determination unit uses, as input, the image data projection-transformed by the projection transformation unit and the temperature distribution data, and uses a learned model learned with the normal state and abnormal state of the furnace wall surface as output to perform an abnormality determination.

[0012] (7) The furnace wall abnormality diagnosis system according to the present invention, in any one of the inventions of (2) to (6) above, uses, as input data, the image data projection-transformed by the projection transformation unit, the temperature distribution data, and the operation data of the coke oven, and uses a learned model machine-learned with a plurality of items causing the temperature abnormality in the carbonization chamber as output data, and further includes an abnormality cause specifying unit for specifying the cause of the abnormality on the furnace wall surface.

[0013] (8) The furnace wall abnormality diagnosis system according to the present invention, in any one of the inventions of (2) to (6) above, uses, as input data, the image data projection-transformed by the projection transformation unit, and uses a learned model machine-learned with a plurality of items causing the temperature abnormality in the carbonization chamber as output data, and further includes an abnormality cause specifying unit for specifying the cause of the abnormality on the furnace wall surface.

[0014] (9) The operation method of the coke oven according to the present invention performs a repair process on the furnace wall of the carbonization chamber based on the specified result of the abnormality cause by the abnormality cause specifying unit provided in the furnace wall abnormality diagnosis system of the invention of (7) or (8) above, and continues the operation.

[0015] (10) The method for repairing a coke oven according to the present invention is characterized in that the repair process of the furnace wall of the carbonization chamber is performed based on the result of specifying the cause of the abnormality by the abnormality cause specifying unit provided in the furnace wall abnormality diagnosis system of the invention of (7) or (8) above.

Effect of the Invention

[0016] The furnace wall temperature distribution measurement system, furnace wall abnormality diagnosis system, operation method of a coke oven, and repair method of a coke oven according to the present invention have the effect of reducing the failure frequency of measuring instruments, enabling the measurement of the temperature distribution of the entire furnace wall surface, and being able to feed back to operation and repair actions by specifying abnormal locations on the furnace wall.

Brief Description of the Drawings

[0017]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Embodiments for Carrying Out the Invention

[0018] Embodiments of the furnace wall temperature distribution measurement system, furnace wall abnormality diagnosis system, operation method of the coke oven, and repair method of the coke oven according to the present invention will be described below. Note that the present invention is not limited by the present embodiment.

[0019] FIG. 1 is a diagram showing the schematic configuration of the furnace wall abnormality diagnosis system 1 according to the embodiment. Note that the furnace wall abnormality diagnosis system 1 according to the embodiment also has a function as a furnace wall temperature distribution measurement system.

[0020] As shown in FIG. 1, the furnace wall abnormality diagnosis system 1 according to the embodiment includes an extruder 2, a coke oven 3, thermal imaging measuring devices 4a and 4b, and an analysis device 5. The extruder 2 includes an extruder main body 20, an extrusion ram 21, and a ram head 22. A driving device for operating the extrusion ram 21 is provided in the extruder main body 20. The ram head 22 is provided at the tip of the extrusion ram 21 on the coke oven 3 side.

[0021] The coke oven 3 for producing coke by carbonizing coal has a structure in which a regenerator is provided at the lower part, and a carbonization chamber (kiln) 31 and a combustion chamber 32 partitioned by furnace wall bricks, which are refractory bricks, are alternately arranged at the upper part. The coal charged into the carbonization chamber 31 is carbonized by heat transfer from the combustion chambers 32 arranged on both sides of the carbonization chamber 31 to become coke. The coke after carbonization is pushed from the extruder 2 side (PS: pusher side) in the furnace length direction by the extrusion ram 21 of the extruder 2 and discharged to the opposite side (CS: coke side). In the coke oven 3 according to the embodiment, a series of operations including coal charging, coking, and coke extrusion (kiln discharging) are repeated.

[0022] In the extruder 2, regardless of the extrusion operation of the coke by the extrusion ram 21, two thermal imaging measuring devices 4a and 4b are always installed on the extruder body 20 located outside the carbonization chamber 31. In the following description, when the two thermal imaging measuring devices 4a and 4b are not particularly distinguished, they may be simply referred to as the thermal imaging measuring device 4. The two thermal imaging measuring devices 4a and 4b are, for example, two-color radiation thermometers having the function of an imaging device such as a camera. The thermal imaging measuring device 4a photographs the entire furnace wall surface of one of the furnace wall surfaces 33a and 33b facing each other in the carbonization chamber 31, and transmits the image data of the photographed furnace wall surface 33a to the analysis device 5 by wireless communication or wired communication. Further, the thermal imaging measuring device 4b photographs the entire furnace wall surface of the other furnace wall surface 33b of the carbonization chamber 31, and transmits the image data of the photographed furnace wall surface 33b to the analysis device 5 by wireless communication or wired communication. In the following description, when the furnace wall surfaces 33a and 33b facing each other in the carbonization chamber 31 are not particularly distinguished, they may be simply referred to as the furnace wall surface 33.

[0023] In the present embodiment, two thermal imaging measuring devices 4a and 4b are provided corresponding to the furnace wall surfaces 33a and 33b of one carbonization chamber 31, respectively. However, the number of thermal imaging measuring devices 4 provided corresponding to one carbonization chamber 31 is not limited to two. Needless to say, the thermal imaging measuring device 4 is installed on the extruder body 20 of the extruder 2 corresponding to each of the plurality of carbonization chambers 31 arranged in the coke oven 3.

[0024] Next, the installation position of the thermal imaging measuring device 4 will be described. Fig. 2(a) is a diagram showing an area A1 that can photograph from the vicinity of the extrusion ram 21 in the furnace length direction to the coke side kiln mouth 35 on the furnace wall surface 33a. Fig. 2(b) is a diagram showing an area A2 that can photograph from the vicinity of the extrusion ram 21 in the furnace length direction to the pusher side kiln mouth 34 on the furnace wall surface 33a. Fig. 2(c) is a diagram showing an area A3 that can photograph from the coke side kiln mouth 35 to the pusher side kiln mouth 34 in the furnace length direction on the furnace wall surface 33a.

[0025] First, as shown in Fig. 2(a), a straight line extending from the position P1 of the coke side opening 35 on one furnace wall surface 33a of the coke oven chamber 31 through the position P2 of the pusher side opening 34 on the other furnace wall surface 33b of the coke oven chamber 31 is defined as the extension line L1. Then, a triangle formed by this extension line L1 and the moving line L2 of the pushing ram 21 (ram head 22) is determined. Thereby, in the furnace wall surface 33a, the region A1 that can be photographed by the thermal imaging measuring instrument 4a can be calculated from the vicinity of the pushing ram 21 to the coke side opening 35 in the furnace length direction.

[0026] Next, as shown in Fig. 2(b), a straight line extending from the position P3 of the pusher side (PS) opening 34 on one furnace wall surface 33a of the coke oven chamber 31 through the tip of the ram head 22 is defined as the extension line L3. Then, a triangle formed by this extension line L3 and the straight line L4 passing through the position P3 of the pusher side opening 34 on the furnace wall surface 33a to the position P2 of the pusher side (PS) opening 34 on the furnace wall surface 33b is determined. Thereby, in the furnace wall surface 33a, the region A2 that can be photographed by the thermal imaging measuring instrument 4a can be calculated from the vicinity of the pushing ram 21 (ram head 22) to the pusher side (PS) opening 34 in the furnace length direction.

[0027] Finally, as shown in Fig. 2(c), the triangle of the overlapping part is obtained by overlapping the region A shown in Fig. 2(a) and the region B shown in Fig. 2(b). Thereby, in the furnace wall surface 33a, the region A3 that can be photographed by the thermal imaging measuring instrument 4a can be calculated from the coke side opening 35 to the pusher side opening 34 in the furnace length direction.

[0028] Incidentally, for the final positioning of the thermal imaging device 4a, in order to avoid the radiant heat from the furnace as much as possible, it is preferable to set the position within the triangular region A3 that is farthest from the pusher-side furnace opening 34. That is, among the three vertices of the triangle forming the region A3 in the top view, it is preferable to arrange the thermal imaging device 4 such that a lens or an imaging element is located at the vertex farthest from the carbonization chamber 31. At this time, as a point to note, the angle of the angle forming the vertex where the thermal imaging device 4a is positioned, in other words, the imaging range angle θ which is the angle for shooting inside the carbonization chamber 31 from the installation position of the thermal imaging device 4a, is smaller than the angle of view of the lens of the thermal imaging device 4a. In the furnace wall abnormality diagnosis system 1 according to the embodiment, for example, with respect to the imaging range angle θ = 5.348 [°], the horizontal angle of view of the lens of the thermal imaging device 4a is set to 33 [°]. Further, the angle formed between the center line of the angle of view of the thermal imaging device 4a and the furnace wall (hereinafter referred to as the imaging angle) is installed in the range of about 40 [°] to 20 [°].

[0029] By the above procedure, the installation position of the thermal imaging device 4a capable of photographing the entire area of one furnace wall surface 33a in the carbonization chamber 31 can be determined. Similarly, for the other furnace wall surface 33b in the carbonization chamber 31, by corresponding the procedure shown in FIGS. 2(a) to 2(c) to the furnace wall surface 33b, the installation position of the thermal imaging device 4b capable of photographing the entire area of the furnace wall surface 33b can be determined.

[0030] Next, the projective transformation of the image data of the furnace wall surface 33 photographed by the thermal imaging device 4 will be described. FIG. 3(a) is a diagram showing an acute-angle photographed image 61 of the furnace wall surface 33 photographed by the thermal imaging device 4. FIG. 3(b) is a diagram showing the planar image 62 after projective transformation. FIG. 3(c) is a diagram showing the temperature distribution data 63.

[0031] The image of the furnace wall surface 33 taken at an acute angle by the thermal imaging device 4 becomes an acute-angle captured image 61 with perspective (foreshortening) as shown in Fig. 3(a). Since correction in the furnace length direction (depth direction) of the furnace wall surface 33 is required, it is somewhat inconvenient. Therefore, projective transformation is performed to flatten the acute-angle captured image 61 taken by the thermal imaging device 4. In this embodiment, the projective transformation is implemented by the software of the analysis device 5 so that the temperature at a representative level (any height position in the furnace height direction of the furnace wall surface 33) can be output from the projectively transformed image. Specifically, as shown in Fig. 3(a), the analysis device 5 designates and sets a rectangular region A11, which is the surface portion to be projectively transformed, from the image of the furnace wall surface 33 (acute-angle captured image 61) taken by the thermal imaging device 4, with four points P11, P12, P13, and P14 that are the vertices of the rectangle. Although the positions of the regions to be projectively transformed may be different for each of the plurality of carbonization chambers 31 arranged in the coke oven 3, it may be a uniform correction in this embodiment. However, the projective transformation may be defined for each carbonization chamber 31 so that the positions after the projective transformation are the same. Then, the analysis device 5 projectively transforms the image portion of the set region A11 and flattens it into a rectangular plane as shown in Fig. 3(b). Next, the planar image 62 that has been projectively transformed and flattened into a rectangular plane is converted into temperature distribution data 63 by a two-color method as shown in Fig. 3(c). Then, the analysis device 5 divides the converted temperature distribution data 63 into 630 points in the furnace length direction × 400 points in the furnace height direction to obtain the temperature at each point. Note that the number of points of the temperature distribution data shown here is only an example, and various numbers of points can be taken in both the furnace length direction and the furnace height direction according to the resolution of the measuring instrument.

[0032] In this way, in the furnace wall abnormality diagnosis system 1 according to the embodiment, the entire furnace wall surface 33 is photographed by the thermal imaging device 4 from outside the carbonization chamber 31. As a result, it is possible to measure the temperature distribution of the entire furnace wall surface 33 while suppressing the thermal imaging device 4 from being exposed to high temperatures and reducing the failure frequency. And in the furnace wall abnormality diagnosis system 1 according to the embodiment, by monitoring the temperature distribution data of the furnace wall surface 33 of the carbonization chamber 31, it is possible to detect temperature abnormalities in a specific carbonization chamber 31 at an early stage. As a result, it is possible to identify factors causing temperature abnormalities, such as blockage of the gas ports in the combustion chamber 32 and excess or deficiency of the gas and air amounts, and lead to early operation actions such as adjustment of the slide damper.

[0033] FIG. 4 is a diagram showing a schematic configuration of the analysis device 5. The analysis device 5 is an electronic computer such as a personal computer, a workstation, and a server, for example. The analysis device 5 includes a control unit 50, an input unit 51, a display unit 52, a storage unit 53, a communication unit 54, and the like. The control unit 50 is constituted by an arithmetic processing device such as a CPU (Central Processing Unit). The control unit 50 has an image acquisition unit 501, a projective conversion unit 502, a temperature conversion unit 503, an abnormality determination unit 504, an abnormality cause identification unit 505, and the like. The image acquisition unit 501 acquires the image data of the furnace wall surface 33 captured by the thermal imaging device 4. The projective conversion unit 502 performs projective conversion on the acquired image data of the furnace wall surface 33. The temperature conversion unit 503 converts the projective-converted image data into temperature distribution data by eliminating the influence of the emissivity caused by the imaging angle. The abnormality determination unit 504 determines the abnormality of the furnace wall surface 33 based on the projective-converted image data of the furnace wall surface 33 and / or the temperature distribution data of the furnace wall surface 33 and / or the operating conditions of the coke (such as the raw material mixing ratio). The abnormality cause identification unit 505 identifies the cause of the abnormality of the furnace wall surface 33. The input unit 51 includes an arbitrary input device capable of detecting the operation of an operator, such as a keyboard, a tablet, a touch pad, and a mouse, and receives an operation related to an instruction for various processes to the analysis device 5. The display unit 52 is a display device such as a monitor. The storage unit 53 includes, for example, a memory, a hard disk drive, a solid state drive, and an optical disk drive, and is a device that stores information necessary for the analysis by the analysis device 5. The communication unit 54 is constituted by a communication interface that communicates via a wired or wireless communication network, and the like.

[0034] Here, when measuring temperature from the radiation amount such as thermography, it is very important to know the emissivity. However, in this embodiment, for the purpose of reducing the heat load, the thermal imaging device 4 is installed on the extruder 2. Along with this, when photographing the furnace wall inside the carbonization chamber 31 from above the extruder 2, as shown in Fig. 2(c), the photographing range angle θ becomes an acute angle. Here, in terms of the positional relationship with the measurement target, if the angle formed between the vertical direction of the measurement target surface and the center line of the angle of view of the thermal imaging device 4 is within 50[°], the measured value can be considered to be within a certain range of variation in emissivity. On the other hand, when the angle formed between the vertical direction of the measurement target surface and the center line of the angle of view of the thermal imaging device 4 exceeds 50[°], the emissivity changes rapidly, and the reliability of the measured value significantly decreases. Therefore, a brightness thermometer or a radiation thermometer that measures the intensity of radiation from the subject must correctly grasp the positional relationship with the measurement target and appropriately set the emissivity at that time. In this embodiment, the photographing angle is an acute angle as described above. This photographing angle, when viewed as the angle formed between the vertical direction of the measurement target surface and the center line of the angle of view of the thermal imaging device 4, is in a state of exceeding 50[°], so adjustment of the emissivity is required as described above. Also, it is not easy to appropriately set the emissivity for the furnace wall that changes day by day due to carbon adhesion and the like. In addition, in addition to the photographing angle being an acute angle in the range of 40[°] or less, since the entire furnace wall is photographed obliquely from the side, the luminance difference between the furnace wall surface 33 close to the thermal imaging device 4 and the furnace wall surface 33 far away also increases, so it is strongly affected by emissivity fluctuations.

[0035] Therefore, in the furnace wall abnormality diagnosis system 1 according to the embodiment, a thermal imaging device 4 using the two-color temperature method is adopted. The two-color temperature method can automatically eliminate the influence of the emissivity caused by the shooting angle by selecting a wavelength band in which the emissivities of two wavelengths of the substance to be measured are the same. At this time, the ratio of the luminances of the measurement targets of the two wavelengths is obtained, compared with the ratio of the luminances of the black body measured in advance, and the temperature of the black body at which the values of both are equal is obtained as the temperature of the measurement target. Therefore, in the two-color temperature method, if the emissivity of the object is the same, it is canceled, so it is possible to measure the temperature of the object without being affected by the emissivity variation without knowing the emissivity itself. In this way, by installing the thermal imaging device 4 outside the carbonization chamber 31 at a position where the entire furnace wall surface of the carbonization chamber 31 can be photographed and utilizing the two-color temperature method, it becomes possible to measure the temperature distribution of the furnace wall as an accurate surface. Also, since the influence of the transmittance is not received by the same theory, temperature measurement is possible even if a substance with the same transmittance such as glass is interposed between the measurement target and the measuring device. Therefore, even when the thermal imaging device 4 is placed in a case having a glass window for photographing, it is possible to measure the temperature of the furnace wall through the glass window, and heat insulation measures, etc., which were difficult with conventional radiation thermometers, become possible. In this way, the thermal imaging device 4 using the two-color temperature method can eliminate not only the influence of the emissivity due to acute-angle shooting but also the influence of the transmittance, so it has advantages in various aspects compared with conventional radiation thermometers.

[0036] FIG. 5 is a flowchart showing an example of the control executed in the furnace wall abnormality diagnosis system 1 according to the embodiment. Here, the control executed after the coke is pushed out from the inside of the carbonization chamber 31 by the push-out ram 21 of the extruder 2 and then the push-out ram 21 is moved so that the ram head 22 is located at the standby position SP will be described.

[0037] First, in the furnace wall abnormality diagnosis system 1, the furnace wall surface 33 of the carbonization chamber 31 is photographed by the thermal imaging device 4 (step S1). Next, in the furnace wall abnormality diagnosis system 1, the image data of the furnace wall surface 33 photographed by the thermal imaging device 4 is transmitted to the analysis device 5 (step S2). Next, in the furnace wall abnormality diagnosis system 1, the analysis device 5 performs projective transformation on the image data of the furnace wall surface 33 to obtain a planar image flattened into a rectangular plane (step S3). Next, in the furnace wall abnormality diagnosis system 1, the analysis device 5 converts the planar image into temperature distribution data using a two-color method (step S4). Next, in the furnace wall abnormality diagnosis system 1, in addition to the temperature distribution data, the analysis device 5 performs an abnormality determination of the furnace wall surface 33 of the carbonization chamber 31 based on the planar image data, the operating conditions of the coke, etc. obtained from the process control computer, etc. (step S5). Next, in the furnace wall abnormality diagnosis system 1, the determination result of the abnormality of the furnace wall surface 33 of the carbonization chamber 31 determined by the analysis device 5 is displayed on the display unit 52 provided in the analysis device 5 (step S6). Then, the furnace wall abnormality diagnosis system 1 ends a series of controls.

[0038] Here, as the furnace wall state in which an abnormality occurs in the furnace wall surface 33 of the carbonization chamber 31, for example, is carbon deposition on the furnace wall surface 33 typified by the cause of clogging. Conventionally, the carbon deposition on the furnace wall surface 33 has mainly been qualitative evaluation and management by visual confirmation by an operator. When carbon adheres to the furnace wall surface 33 of the carbonization chamber 31, a carbon layer is formed. Considering heat transfer from the combustion chamber 32 to the carbonization chamber 31, the heat transfer distance becomes longer by the thickness of the carbon layer rather than just the thickness of the furnace wall. Therefore, on the furnace wall surface 33 of the carbonization chamber 31, the temperature of the part with the carbon layer tends to be measured lower than the temperature of the part without the carbon layer. Also, in the carbonization chamber 31 of the aged coke oven 3, damage to the furnace wall surface 33 such as bricks being chipped becomes noticeable. In this case, the heat transfer distance from the combustion chamber 32 to the carbonization chamber 31 becomes shorter by the amount of the chipped bricks. Therefore, on the furnace wall surface 33 of the carbonization chamber 31, the temperature of the part where the bricks are chipped tends to be measured higher than the temperature of the part where the bricks are not chipped. In any case, a change occurs in the temperature distribution data.

[0039] Therefore, for the temperature distribution data of the furnace wall surface 33, one or more threshold values may be set to determine and output the abnormal level. For example, the analysis device 5 may determine that a portion of the furnace wall surface 33 where the temperature exceeds a predetermined first threshold value or is lower than a second threshold value is abnormal, or may set a plurality of high-temperature levels and low-temperature levels for determination. Further, the combination of each abnormal level and the abnormal occurrence position can be used to specify the abnormal location of the furnace wall and estimate the abnormal level. Then, by using the estimated result as one of the indicators for determining the priority order of the carbonization chambers 31 to be repaired in the coke oven 3, it is possible to prevent troubles that may occur in the carbonization chambers 31 due to the abnormality of the furnace wall in advance.

[0040] Further, the analysis device 5 may input the temperature distribution data of the projected furnace wall surface 33 and perform an abnormality determination of the furnace wall surface 33 using a learned model that has been machine-learned with the normal state and the abnormal state of the furnace wall surface 33 as outputs.

[0041] Regarding the learning method, for example, temperature distribution data including abnormal locations is used to identify the abnormal locations and learn the abnormal patterns. For example, the abnormal part of the temperature data can be surrounded by a rectangle or the like and specified and extracted, and the learning data can be prepared by labeling normal and abnormal.

[0042] A learned model is constructed using such learning data, and using the obtained learned model, temperature distribution data is input to the model learned by the temperature distribution data, and as the output of the learned model, the determination results of normal and abnormal of the furnace wall surface 33 can be output. Also, for the abnormality determination, the position information can also be output. Further, separately from the output of the learned model, it may be possible to output the presence or absence of an abnormality with respect to the operation temperature control value and the separation temperature from the normal state.

[0043] Note that, here, the input of the abnormality determination model was learned using only the temperature distribution data of the furnace wall surface 33. However, a learning model may be constructed to output the presence or absence of an abnormality by using the operation data of the coke oven 3 at the time of acquiring the temperature distribution data as input data together. Examples of the operation data of the coke oven 3 include raw material information such as coal moisture content and raw material blending ratio, and the flue temperature of the combustion chamber 32 during operation.

[0044] The analysis device 5 may store the image data of the furnace wall surface 33 of the carbonization chamber 31 photographed by the thermal imaging device 4 in the storage device, and perform analysis by machine learning such as deep learning with the image data of the furnace wall surface 33 as input and the abnormality of the furnace wall surface 33 as output. Thereby, in the furnace wall abnormality diagnosis system 1 of the coke oven 3 according to the embodiment, the monitoring of the furnace wall state of the carbonization chamber 31 can be further strengthened. Here, the image data of the furnace wall surface 33 refers to the original image data captured by the thermal imaging device 4, which has been subjected to projective transformation and refers to the image data before being converted into temperature distribution data.

[0045] When trying to determine the abnormality of the furnace wall using only the temperature distribution data, the furnace wall temperature may be high or low regardless of carbon deposition or brick damage. Therefore, it is assumed that it is less reliable to determine the abnormality of the furnace wall surface 33 caused by carbon deposition or brick damage based only on the temperature distribution data, so image data is used.

[0046] The image data of the furnace wall surface 33 photographed by the thermal imaging device 4 is thermal image data and is an image in the infrared region. However, the image data before being converted into temperature distribution data has a low luminance level and can be said to be data capable of discriminating the information of the furnace wall shape. Therefore, in the furnace wall abnormality diagnosis system 1 of the coke oven 3 according to the embodiment, the image data of the furnace wall surface 33 that has been subjected to projective transformation can be used as the input of machine learning.

[0047] Regarding the learning method, for example, in the same way as learning a learning model using temperature distribution data, using the image data of the furnace wall surface 33 including abnormal parts, identify the abnormal parts and learn the abnormal patterns. For example, the abnormal part of the image data can be surrounded by a rectangle or the like and specified and extracted, and the learning data can be prepared by labeling normal and abnormal.

[0048] Using such learning data, a learned model is constructed, and using the obtained learned model, input the image data into the model learned by the image data, and as the output of the learned model, output the determination result of normal and abnormal of the furnace wall surface 33. Also, for the abnormality determination, the position information can also be output.

[0049] Note that here, the input of the abnormality determination model was learned using only the image data of the furnace wall surface 33, but a learning model may be constructed to use the operation data of the coke oven 3 at the time of image data acquisition as input data together to output the presence or absence of abnormality. Examples of the operation data of the coke oven 3 include raw material information such as coal moisture content and raw material mixing ratio, and the flue temperature of the combustion chamber 32 during operation.

[0050] Also, regarding the abnormality determination method using image data, a method of registering a plurality of abnormal shape information and abnormal size information in advance as templates and determining the abnormality level from the matching ones can also be applied.

[0051] Note that for the image data, the learning data may be increased by randomly changing the brightness level and orientation within a predetermined range.

[0052] Then, the analysis device 5 may output the result of determining the temperature abnormality of the furnace wall surface 33 from the temperature distribution data and the result of determining the shape abnormality of the furnace wall from the image data of the furnace wall surface 33. Further, the analysis device 5 can link these two determination results, combine the two determination conditions, and specify and output the abnormal content corresponding to the abnormal location and cause of the furnace wall. Specifically, various methods can be adopted, such as determining the abnormality by the AND condition of these two determination results, outputting a warning by the OR condition, or making an abnormality determination in consideration of other operating conditions.

[0053] The analysis device 5 may further use a learned model that is machine-learned with the input being the image data of the furnace wall surface 33 after projective transformation and the temperature distribution data of the furnace wall surface 33, and the output being the normal state and the abnormal state of the furnace wall surface 33 to perform an abnormality determination of the furnace wall surface 33. It can also be configured to further output the position where the abnormality occurs as the output. As the machine learning method, for example, deep learning may be used, or other learning models may be used.

[0054] As another embodiment, as input data, in addition to the temperature distribution data of the furnace wall surface 33 and the image data of the furnace wall surface 33, the operating conditions of the coke oven 3 can be obtained from the host computer and used. The operating conditions include, in addition to the coal moisture content and the blending ratio of the raw materials, the flue temperature of the combustion chamber 32, etc. By including at least one of these, the operating state can be reflected and the accuracy of the abnormality determination can be improved. Therefore, a learned model for determining the normality / abnormality of the furnace wall surface 33 and specifying the abnormal location can be constructed and used with all or a partial combination of the temperature distribution data of the furnace wall surface 33, the image data of the furnace wall surface 33, and the operating conditions of the coke oven.

[0055] FIG. 6 is a diagram showing an outline of the abnormality monitoring function in the furnace wall abnormality diagnosis system 1 according to the embodiment.

[0056] In the furnace wall abnormality diagnosis system 1 according to the embodiment, the analysis device 5 may further include an abnormality cause specifying unit 505 that specifies the cause of the abnormality on the furnace wall surface 33. When performing abnormal monitoring of the carbonization chamber 31, first, the analysis device 5 generates a new abnormality determination model for determining the abnormality of the furnace wall and specifying the cause of the abnormality by machine learning. For example, using the learning data 74 including temperature distribution data and image data corresponding to the normal state and abnormal state of the furnace wall surface 33, and other operation data of the coke oven 3 as input, an abnormality determination model 73 is generated by machine learning with the normality / abnormality of the furnace wall and a plurality of items that cause abnormalities in the carbonization chamber 31 as output. Note that the other operation data includes raw material information such as coal moisture content, blending ratio of raw materials, and flue temperature in the combustion chamber 32. In addition, the plurality of items that cause abnormalities in the carbonization chamber 31 include abnormality determination data 75 and abnormality content data 76. The abnormality determination data 75 includes the abnormality determination result of the furnace wall and the abnormal location on the furnace wall surface 33. The abnormality content data 76 includes items such as carbonization failure, clearance reduction, coke fragmentation, carbon deposition, and brick damage as furnace internal abnormalities. Here, these abnormal contents indicate the causes of the furnace internal abnormalities. For creating a learning model that outputs the abnormal contents, according to the abnormal contents (such as carbonization failure, clearance reduction, coke fragmentation, carbon deposition, and brick damage), the abnormal position on the furnace wall surface 33 may be specified by a closed space such as a rectangle, extracted, and then categorized and labeled according to the abnormal contents.

[0057] In creating a learning model that outputs the abnormal contents, according to the abnormal contents (such as carbonization failure, clearance reduction, coke fragmentation, carbon deposition, and brick damage), the abnormal position on the furnace wall surface 33 is specified by a closed space such as a rectangle, and the temperature data and image data of the abnormal position on the furnace wall surface 33 are extracted. Also, the operation data of the coke oven 3 associated with the data including the abnormal locations of each furnace wall surface 33 is acquired. Then, it is categorized and labeled according to the abnormal contents to create learning data.

[0058] Then, using as input any one of the temperature data, image data of the furnace wall surface 33, and the operation data of the coke oven 3 associated with these data, or a combination thereof, a learning model that outputs the normality / abnormality of the furnace wall surface 33 can be generated by machine learning. Further, in the case of an abnormality, a learning model can also be generated so as to output which category of the abnormality content it belongs to.

[0059] Then, in the analysis device 5, the furnace wall data 71 and other operation data 72 are input into an abnormality determination model 73 which is a learned model. Note that the furnace wall data 71 includes the image data of the projected furnace wall surface 33 and the temperature distribution data of the furnace wall surface 33. Thereby, in the analysis device 5, abnormality determination data 75 which is the first output data and abnormality content data 76 which is the second output data are output. Note that the abnormality determination data 75 which is the first output data includes the abnormality determination result of the furnace wall and information regarding the abnormal location of the furnace wall surface 33. Further, the abnormality content data 76 which is the second output data includes furnace temperature abnormalities (carbonization failure, clearance reduction, coke fragmentation, and carbon deposition).

[0060] In this way, in the furnace wall abnormality diagnosis system 1 according to the embodiment, by generating an abnormality determination model that determines the details of the abnormality cause of the furnace wall, it becomes possible for the analysis device 5 to also estimate abnormalities that cannot be judged from the captured images such as carbonization failure and coke fragmentation. Thereby, when these troubles occur, the details of the trouble can be estimated early by the analysis device 5 and linked to repair actions such as thermal spraying repair of the bricks on the furnace wall. Therefore, as an operation method of the coke oven 3 according to the embodiment, by performing a repair process on the furnace wall of the carbonization chamber 31 by a repair method of the coke oven 3 based on the specific result of the abnormality cause of the furnace wall surface 33 by the analysis device 5, smooth operation of the coke oven 3 can be continued.

Example

[0061] FIG. 7 is a diagram showing the temperature distribution in the furnace length direction at a specific furnace height position in the temperature distribution of the furnace wall surface 33 measured using the furnace wall abnormality diagnosis system 1 according to the embodiment. The temperature distribution of the furnace wall surface 33 shown in FIG. 7 was measured by the thermal imaging device 4 after the ram head 22 returned to the standby position SP after the coke was pushed out from the carbonization chamber 31 by the pushing ram 21.

[0062] In the first embodiment, using the furnace wall abnormality diagnosis system 1 according to the embodiment, the temperature distributions from the pusher-side kiln mouth 34 to the vicinity of the center in the furnace length direction at three representative levels (any height position in the furnace height direction of the furnace wall surface 33), namely, the upper, middle, and lower levels in the furnace height direction of the furnace wall surface 33, were obtained. In FIG. 7, the left end of the horizontal axis indicates the end on the pusher-side kiln mouth 34 side, and the right direction of the horizontal axis indicates the inner direction of the furnace. Note that "Lev1" in FIG. 7 is a graph of the representative level showing the temperature distribution in the furnace length direction at an arbitrary height position in the upper part in the furnace height direction of the furnace wall surface 33. "Lev2" in FIG. 7 is a graph of the representative level showing the temperature distribution in the furnace length direction at an arbitrary height position in the middle part in the furnace height direction of the furnace wall surface 33. "Lev3" in FIG. 7 is a graph of the representative level showing the temperature distribution in the furnace length direction at an arbitrary height position in the lower part in the furnace height direction of the furnace wall surface 33.

[0063] From FIG. 7, it can be confirmed that at any representative level, the temperature is low on the kiln mouth 34 side, and the temperature increases toward the center in the furnace length direction of the furnace wall surface 33. This is the same behavior as the temperature of the furnace wall surface 33 on the kiln mouth 34 side that is in contact with the atmosphere being low. Also, when the correlation between the temperature measurement results in the furnace length direction of the existing combustion chamber 32 and the representative values was taken, a strong positive correlation with a correlation coefficient of 0.82 was confirmed. From these facts, it can be seen that in the furnace wall abnormality diagnosis system 1 of the coke oven 3 according to the embodiment, a reasonable temperature distribution on the furnace wall surface 33 of the carbonization chamber 31 can be captured both physically and from the comparison with the known furnace internal temperature distribution tendency.

[0064] Thus, it can be understood that even by using the thermal imaging device 4 installed outside the furnace, it is possible to measure the temperature of the entire furnace wall in the furnace length direction without being affected by the emissivity variation. Further, by extracting a plurality of temperatures in the furnace height direction, it is also possible to extract the temperature distribution of the furnace wall surface 33. It can be understood that the installation position of the thermal imaging device 4 is also a position where measurement can be performed from the outside of the carbonization chamber 31, which reduces the failure frequency of the thermal imaging device 4 and enables measurement of the temperature distribution of the entire furnace wall surface. Furthermore, since temperature measurement is possible and image information of the furnace wall can also be obtained, it is possible to identify abnormal portions of the furnace wall and feedback them to operation and repair actions.

Example

[0065] In this Example 2, using the furnace wall abnormality diagnosis system 1 according to the embodiment, the temperature data of the furnace wall surface 33 of the carbonization chamber 31 and the data of the flue temperature of the combustion chamber 32 measured by existing equipment as the operating conditions of the coke oven 3 are used to show the result of constructing an abnormality determination model.

[0066] As the input of the abnormality determination model, among the temperature distribution data changed from the thermal image of the furnace wall surface 33 to temperature, the temperature data obtained in association with the flue position through which the combustion gas in the combustion chamber 32 passes is used. Therefore, 15 points of temperature data associated with the flue position between the kiln mouth and the kiln center in the furnace length direction of one side furnace wall surface are used as the temperature data, and 15 points of the flue temperature of the combustion chamber 32 are used as the input. Note that the number of measurement points for each is not limited to this number, and it is not necessarily the same. Temperature measurement may be performed within the range of the corresponding positions of the carbonization chamber 31 and the combustion chamber 32. Also, as learning data, data determined as a normal kiln by the operator is used as normal data, and data immediately after a clogging occurs due to a temperature abnormality is used as abnormal data, and classification of normal / abnormal is performed. The number of data points used is 937 points of normal data (647 points for learning, 290 points for verification) and 182 points of abnormal data (133 points for learning, 49 points for verification). The accuracy verification result of the learning model is shown in Table 1.

[0067]

Table 1

[0068] The estimation accuracy of normal and abnormal conditions is a correct rate of about 97% ((290 + 39) / (290 + 10 + 39) = 0.9705), and a result that can be used as an abnormal determination model for clogging was obtained.

[0069] Thus, it can be seen that an abnormal determination model for clogging can be constructed by an abnormal determination model that uses the flue temperature among the temperature data and operation data of the furnace wall surface 33 as input data. Also, if the image information of the furnace wall is used, the abnormal location of the furnace wall can be specified and feedback can be provided to operation and repair actions.

Industrial Applicability

[0070] The present invention can provide a furnace wall temperature distribution measurement system, a furnace wall abnormality diagnosis system, an operation method of a coke oven, and a repair method of a coke oven that can reduce the failure frequency of measuring instruments and measure the temperature distribution of the entire furnace wall surface.

Explanation of Signs

[0071] 1 Furnace wall abnormality diagnosis system 2 Extruder 3 Coke oven 4, 4a, 4b Thermal imaging device 5 Analysis device 20 Extruder main body 21 Extrusion ram 22 Ram head 31 Carbonization chamber 32 Combustion chamber 33, 33a, 33b Furnace wall surface 34 Kiln mouth 35 Kiln mouth 50 Control unit 51 Input unit 52 Display unit 53 Storage unit 54 Communication unit 61 Acute angle photographed image 62 Planar image 63 Temperature distribution data 71 Furnace wall data 72 Other operating data 73 Abnormality determination model 74 Learning data 75 Abnormality determination data 76 Abnormality content data 501 Image acquisition unit 502 Projection conversion unit 503 Temperature conversion unit 504 Abnormality determination unit 505 Abnormality cause identification unit

Claims

1. A thermal imaging measuring device installed at a position capable of imaging the entire furnace wall surface of a carbonization chamber disposed in a coke oven on an extruder, an image acquisition unit that acquires image data of the furnace wall surface obtained by imaging the entire furnace wall surface with the thermal imaging measuring device, a projective transformation unit that performs projective transformation on the acquired image data of the furnace wall surface, a temperature conversion unit that converts the projected image data into temperature distribution data by eliminating the influence of emissivity caused by the imaging angle by a two-color method, A furnace wall temperature distribution measurement system, characterized by comprising the above.

2. A thermal imaging measuring device installed at a position capable of imaging the entire furnace wall surface of a carbonization chamber disposed in a coke oven on an extruder, an image acquisition unit that acquires image data of the furnace wall surface obtained by imaging the entire furnace wall surface with the thermal imaging measuring device, a projective transformation unit that performs projective transformation on the acquired image data of the furnace wall surface, a temperature conversion unit that converts the projected image data into temperature distribution data by eliminating the influence of emissivity caused by the imaging angle by a two-color method, an abnormality determination unit that determines an abnormality of the furnace wall surface based on the image data subjected to projective transformation by the projective transformation unit and / or the temperature distribution data converted by the temperature conversion unit, A furnace wall abnormality diagnosis system, characterized by comprising the above.

3. The furnace wall abnormality diagnosis system according to claim 2, wherein the abnormality determination unit determines a portion where the temperature exceeds a predetermined first threshold value or a portion lower than a second threshold value as an abnormality.

4. The furnace wall abnormality diagnosis system according to claim 2, wherein the abnormality determination unit registers a plurality of pieces of information on abnormal shapes and abnormal sizes in advance as templates for the image data, and determines an abnormality level from the matching ones.

5. The furnace wall abnormality diagnosis system according to claim 2, wherein the abnormality determination unit uses a first learned model that is machine-learned with the image data subjected to projective transformation by the projective transformation unit as an input and the normal state and abnormal state of the furnace wall surface as outputs, and a second learned model that is machine-learned with the temperature distribution data as an input and the normal state and abnormal state of the furnace wall surface as outputs to perform abnormality determination.

6. The abnormality determination unit uses the image data projected and transformed by the projective transformation unit and the temperature distribution data as inputs, and performs abnormality determination using a learned model that has been machine-learned with the normal state and abnormal state of the furnace wall surface as outputs. The furnace wall abnormality diagnosis system according to claim 2, characterized in that.

7. Using the image data projected and transformed by the projective transformation unit, the temperature distribution data, and the operation data of the coke oven as input data, and a learned model that has been machine-learned with a plurality of items that cause temperature abnormalities in the carbonization chamber as output data, The furnace wall abnormality diagnosis system according to any one of claims 2 to 6, further comprising an abnormality cause specifying unit that specifies the cause of the abnormality on the furnace wall surface.

8. Using the image data projected and transformed by the projective transformation unit as input data, and a learned model that has been machine-learned with a plurality of items that cause temperature abnormalities in the carbonization chamber as output data, The furnace wall abnormality diagnosis system according to any one of claims 2 to 6, further comprising an abnormality cause specifying unit that specifies the cause of the abnormality on the furnace wall surface.

9. Based on the specific result of the abnormality cause by the abnormality cause specifying unit provided in the furnace wall abnormality diagnosis system according to claim 7, a repair process of the furnace wall of the carbonization chamber is performed, and the operation is continued. A method for operating a coke oven, characterized in that.

10. Based on the specific result of the abnormality cause by the abnormality cause specifying unit provided in the furnace wall abnormality diagnosis system according to claim 8, a repair process of the furnace wall of the carbonization chamber is performed, and the operation is continued. A method for operating a coke oven, characterized in that.

11. Based on the specific result of the abnormality cause by the abnormality cause specifying unit provided in the furnace wall abnormality diagnosis system according to claim 7, a repair process of the furnace wall of the carbonization chamber is performed. A method for repairing a coke oven, characterized in that.

12. Based on the specific result of the abnormality cause by the abnormality cause specifying unit provided in the furnace wall abnormality diagnosis system according to claim 8, a repair process of the furnace wall of the carbonization chamber is performed. A method for repairing a coke oven, characterized in that.

Images