Computer program, information processing method, and information processing apparatus

Abstract

This invention provides a computer program, an information processing method, and an information processing device for detailed observation of the inside of a bubbling fluidized bed combustion furnace. [Solution] The computer program acquires images of the inside of a bubbling fluidized bed combustion furnace using an infrared camera, generates a brightness distribution inside the bubbling fluidized bed combustion furnace and a velocity vector distribution representing the distribution of velocity vectors of moving objects inside the bubbling fluidized bed combustion furnace based on the images, and causes the computer to execute a process to output analysis results based on the brightness distribution and the velocity vector distribution.

Description

Technical Field

[0001] The present invention relates to a computer program, an information processing method, and an information processing apparatus for observing the inside of a bubbling fluidized bed combustion furnace.

Background Art

[0002] As a combustion furnace for burning fuels such as biomass or waste, there is a bubbling fluidized bed combustion furnace. In this combustion furnace, air is passed through a fluidized medium filled inside the furnace from below, and a fluidized bed is formed in which the fluidized medium flows violently as if a liquid is boiling. The fuel is introduced into the fluidized bed, comes into contact with the high-temperature fluidized medium and air, and burns. Patent Document 1 discloses a technique for observing the inside of a bubbling fluidized bed combustion furnace using an infrared camera.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In the conventional technology, the temperature distribution inside a bubbling fluidized bed combustion furnace can be obtained by using an infrared camera. However, there is a need to observe the inside of a bubbling fluidized bed combustion furnace in more detail. For example, there is a need to observe the velocity distribution of an object inside a bubbling fluidized bed combustion furnace.

[0005] An object of the present invention is to provide a computer program, an information processing method, and an information processing apparatus for observing the inside of a bubbling fluidized bed combustion furnace in detail.

Means for Solving the Problems

[0006] A computer program according to one embodiment of the present invention is characterized by causing a computer to perform the following processes: acquire an image of the inside of a bubbling fluidized bed combustion furnace taken with an infrared camera; generate a luminance distribution inside the bubbling fluidized bed combustion furnace and a velocity vector distribution representing the distribution of velocity vectors of moving objects inside the bubbling fluidized bed combustion furnace based on the image; and output an analysis result based on the luminance distribution and the velocity vector distribution.

[0007] An information processing method according to one embodiment of the present invention is characterized by acquiring an image of the inside of a bubbling fluidized bed combustion furnace captured by an infrared camera, generating a brightness distribution inside the bubbling fluidized bed combustion furnace and a velocity vector distribution representing the distribution of velocity vectors of moving objects inside the bubbling fluidized bed combustion furnace based on the image, and outputting analysis results based on the brightness distribution and the velocity vector distribution.

[0008] An information processing device according to one embodiment of the present invention comprises a calculation unit, the calculation unit acquires an image of the inside of a bubbling fluidized bed combustion furnace captured by an infrared camera, generates a brightness distribution inside the bubbling fluidized bed combustion furnace and a velocity vector distribution representing the distribution of velocity vectors of moving objects inside the bubbling fluidized bed combustion furnace based on the image, and outputs an analysis result based on the brightness distribution and the velocity vector distribution.

[0009] In one embodiment of the present invention, based on images taken of the inside of a bubbling fluidized bed combustion furnace with an infrared camera, a brightness distribution and a velocity vector distribution are generated inside the bubbling fluidized bed combustion furnace, and analysis results based on the brightness distribution and velocity vector distribution are output. Using the output analysis results, the inside of the bubbling fluidized bed combustion furnace is observed in detail. For example, the state of the bubbling fluidized bed inside the bubbling fluidized bed combustion furnace is observed. [Effects of the Invention]

[0010] The present invention offers excellent advantages, such as enabling detailed observation of the inside of a bubbling fluidized bed combustion furnace. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic diagram showing an example of the configuration of an information processing system. [Figure 2] This is a schematic diagram showing an example of a plant configuration. [Figure 3] This is a block diagram showing an example of the internal configuration of an information processing device. [Figure 4] This flowchart shows an example of the steps an information processing device performs to observe the inside of a combustion furnace. [Figure 5] This figure shows an example of a heatmap and the first graph of the brightness distribution. [Figure 6] This figure shows an example of a heatmap and a second graph of the velocity vector distribution. [Figure 7] This is a conceptual diagram illustrating examples of the functions of the anomaly detection model, the first feature extraction model, and the second feature extraction model. [Figure 8] This flowchart shows an example of the steps an information processing device performs to detect an anomaly inside a combustion furnace. [Figure 9] This flowchart shows an example of a procedure for controlling other devices in response to abnormalities inside a combustion furnace. [Modes for carrying out the invention]

[0012] The present invention will be described in detail below with reference to the drawings illustrating its embodiments. Figure 1 is a schematic diagram showing an example configuration of the information processing system 100. The information processing system 100 is a system that performs processing for observing plant 2. The information processing system 100 includes an information processing device 1. The information processing device 1 is connected to plant 2. For example, the information processing device 1 is connected to plant 2 via a communication network using wireless or wired communication. The information processing device 1 may be connected to multiple plants 2.

[0013] Figure 2 is a schematic diagram showing an example configuration of Plant 2. Plant 2 is a plant that burns fuel such as biomass or waste and utilizes the heat generated. The fuel includes, for example, waste plastics, such as PRF (refuse derived paper and plastics densified fuel). Plant 2 generates electricity using the generated heat. Plant 2 is equipped with a combustion furnace 3 and a heat exchanger 4. The combustion furnace 3 and heat exchanger 4 constitute a boiler. Plant 2 includes other equipment. For example, the power-generating Plant 2 includes piping through which steam generated from the heat exchanger 4 passes, a turbine that rotates with the steam, and a generator connected to the turbine. Furthermore, Plant 2 is equipped with an infrared camera 32 that photographs the inside of the combustion furnace 3. In Figure 2, equipment other than the combustion furnace 3, infrared camera 32, and heat exchanger 4 included in Plant 2 is omitted.

[0014] The combustion furnace 3 is a bubbling fluidized bed combustion furnace that includes a bubbling fluidized bed 31. The combustion furnace 3 is formed in a cylindrical shape overall, for example, a rectangular cylinder. The lower part of the inside of the combustion furnace 3 is filled with a fluidized medium in which fine particles such as sand have accumulated, and air is passed through the fluidized medium from below, forming a bubbling fluidized bed 31 in which the fluidized medium flows vigorously as if a liquid were boiling. The air that is passed through the fluidized medium from below is not limited to air that is passed through the fluidized medium from directly below along the vertical direction, but may also be air that is passed through the fluidized medium from diagonally below along a direction intersecting the vertical direction. The fluidized medium contained in the bubbling fluidized bed 31 is heated. Heat transfer tubes 33 are arranged inside the combustion furnace 3. The heat transfer tubes 33 pass through the bubbling fluidized bed 31. The heat transfer tubes 33 are connected to a boiler which consists of the combustion furnace 3 and a heat exchanger 4. The heat transfer tubes 33 may also be connected to equipment located outside the combustion furnace 3. A fluid flows inside the heat transfer tube 33, absorbing heat from the bubbling fluidized bed 31.

[0015] In the combustion furnace 3, a fuel inlet 34 is provided at a position above the bubbling fluidized bed 31. The fuel introduced into the interior of the combustion furnace 3 from the fuel inlet 34 falls above the bubbling fluidized bed 31 and is thus introduced into the bubbling fluidized bed 31. The fuel introduced into the bubbling fluidized bed 31 is agitated in the bubbling fluidized bed 31, comes into contact with a high-temperature fluid medium and air, and burns. When the fuel burns, a flame 35 is generated and heat is generated. The generated heat is recovered by the combustion furnace 3, the heat transfer pipe 33, and the heat exchanger 4.

[0016] An infrared camera 32 is provided outside the combustion furnace 3. The infrared camera 32 photographs the interior of the combustion furnace 3 from the outside of the combustion furnace 3. The infrared camera 32 performs photography using infrared rays. The infrared camera 32 is disposed at a position above the bubbling fluidized bed 31 and the fuel inlet 34. The infrared camera 32 photographs the interior of the combustion furnace 3 from above. The infrared camera 32 receives light having a wavelength that is not absorbed by combustion gases such as water vapor or carbon dioxide. For this reason, the infrared camera 32 can photograph the bubbling fluidized bed 31 through the flame 35. The infrared camera 32 is connected to the information processing device 1. For example, the infrared camera 32 is connected to the information processing device 1 via a communication network.

[0017] Although not shown in FIG. 2, the plant 2 may include a ventilation device, a fluid medium discharge device, a fluid medium sorting and processing device, a fluid medium circulation device, or a fluid medium replenishment device as devices for properly maintaining the flow state of the fluid medium contained in the bubbling fluidized bed 31. The ventilation device is a device for passing air through the bubbling fluidized bed 31 and includes a device for heating the air. The fluid medium discharge device is a device for discharging the fluid medium from the bubbling fluidized bed 31 and includes a device for cooling the fluid medium. The fluid medium sorting and processing device is a device for maintaining the fluid medium at an appropriate particle size by removing inappropriate particle size substances in the fluid medium and sorting only the appropriate fluid medium. The fluid medium sorting and processing device may perform pulverization of the fluid medium or addition of a chemical agent to the bubbling fluidized bed 31. The fluid medium circulation device is a device for returning the fluid medium, which has been optimized by the fluid medium sorting and processing device after being discharged from the bubbling fluidized bed 31, to the bubbling fluidized bed 31. The fluid medium replenishment device is a device for storing unused fluid medium and replenishing it to the bubbling fluidized bed 31.

[0018] In managing the combustion furnace 3, there is a need to observe the inside of the bubbling fluidized bed 31 in detail. Also, there is a need to properly maintain the flow state and make the consumption of the fluid medium neither excessive nor insufficient by taking improvement actions to improve the state of the bubbling fluidized bed 31 before the flow state of the bubbling fluidized bed 31 deteriorates. The improvement actions are processes such as cleaning the fluid medium, replacing (discharging and replenishing) the fluid medium, or changing the air flow rate. Based on the observation results of the inside of the bubbling fluidized bed 31, control of the improvement actions for the bubbling fluidized bed 31 can be performed. In the present embodiment, the information processing device 1 performs an information processing method for observing the inside of the combustion furnace 3 using the infrared camera 32.

[0019] Figure 3 is a block diagram showing an example of the internal configuration of the information processing device 1. The information processing device 1 is configured using a computer such as a personal computer or a server device. The information processing device 1 includes an arithmetic unit 11, a memory 12, a storage unit 13, a reading unit 14, an operation unit 15, a display unit 16, and an input / output unit 17. The arithmetic unit 11 is configured using, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a multi-core CPU. The arithmetic unit 11 may also be configured using a quantum computer. The memory 12 stores temporary data generated in connection with calculations. The memory 12 is, for example, RAM (Random Access Memory). The storage unit 13 is non-volatile and is, for example, a hard disk or a non-volatile semiconductor memory. The reading unit 14 reads information from a recording medium 10 such as an optical disc or portable memory.

[0020] The operation unit 15 accepts input of information such as text by receiving operations from the user. The operation unit 15 is, for example, a keyboard, a pointing device, or a touch panel. The display unit 16 displays images. The display unit 16 is, for example, a liquid crystal display or an EL display (Electroluminescent Display). The operation unit 15 and the display unit 16 may be integrated. The input / output unit 17 is connected to the plant 2 and performs data input and output to the plant 2. For example, the input / output unit 17 is an interface. The input / output unit 17 is connected to the infrared camera 32, and the information processing device 1 receives data from the infrared camera 32 at the input / output unit 17.

[0021] The arithmetic unit 11 causes the reading unit 14 to read the computer program (program product) 131 recorded on the recording medium 10, and stores the read computer program 131 in the storage unit 13. The arithmetic unit 11 executes processing to realize the functions of the information processing device 1 according to the computer program 131. The computer program 131 may be stored in the storage unit 13 in advance, or it may be downloaded from outside the information processing device 1. In this case, the information processing device 1 does not need to have a reading unit 14.

[0022] The computer program 131 can be deployed on a single computer, located at a single site, or distributed across multiple sites and run on multiple computers interconnected by a communication network. That is, the information processing device 1 may consist of multiple computers, and the computer program 131 may run on multiple computers connected via a communication network. The information processing device 1 may also be configured using a cloud server.

[0023] The processing steps described below for executing the information processing method can be performed on multiple computers. The processing steps can also be performed on different computers. The data used during processing may be stored on multiple computers. The processing steps can also be performed using a virtual machine. The processing steps may be performed by multiple arithmetic units. The processing steps may also be performed by different arithmetic units. For example, part of the processing may be performed on one computer, and other parts on other computers.

[0024] The information processing device 1 executes an information processing method for observing the inside of the combustion furnace 3. Figure 4 is a flowchart showing an example of the procedure for the processing performed by the information processing device 1 to observe the inside of the combustion furnace 3. Hereinafter, steps will be abbreviated as S. The information processing device 1 performs the following processing by having the arithmetic unit 11 perform information processing according to the computer program 131. The information processing device 1 acquires an image of the inside of the combustion furnace 3 taken by the infrared camera 32 (S101). The infrared camera 32 takes pictures of the inside of the combustion furnace 3 while the fuel is burning inside the combustion furnace 3. At this time, the infrared camera 32 takes pictures of the inside of the combustion furnace 3 continuously or intermittently. The infrared camera 32 generates luminance data representing the two-dimensional distribution of infrared intensity inside the combustion furnace 3 by taking pictures, and transmits the luminance data to the information processing device 1.

[0025] In S101, the information processing device 1 receives luminance data transmitted from the infrared camera 32 in the input / output unit 17, and the calculation unit 11 generates an image based on the received luminance data, thereby acquiring an image. Alternatively, in S101, the infrared camera 32 may create an image representing the captured content and transmit the created image to the information processing device 1, and the information processing device 1 may acquire an image by receiving the image transmitted from the infrared camera 32 in the input / output unit 17. The image acquired by the information processing device 1 is a moving image obtained by the infrared camera 32 continuously photographing the inside of the combustion furnace 3. Alternatively, the image acquired by the information processing device 1 is a time-series image consisting of multiple still images obtained by the infrared camera 32 intermittently photographing the inside of the combustion furnace 3. The image acquired by the information processing device 1 is an infrared image representing the intensity distribution of infrared radiation generated inside the combustion furnace 3, and the image shows the bubbling fluidized bed 31. More specifically, the image shows the fluid medium and fuel contained in the bubbling fluidized bed 31. The calculation unit 11 stores the acquired image in the storage unit 13.

[0026] The information processing device 1 then generates a luminance distribution inside the combustion furnace 3 (S102). In S102, the calculation unit 11 generates a luminance distribution inside the combustion furnace 3 based on the acquired image. Luminance corresponds to infrared intensity. The luminance distribution includes data indicating the luminance in each part inside the combustion furnace 3. The luminance distribution may be represented as an image. The luminance distribution includes the luminance of each part of the bubbling fluidized bed 31. For example, the luminance of each part of the bubbling fluidized bed 31 is a value corresponding to the temperature of each part. In S102, the calculation unit 11 generates multiple luminance distributions at multiple points in time based on a moving image or a time-series image. The calculation unit 11 stores the generated luminance distributions in the storage unit 13.

[0027] The information processing device 1 generates a velocity vector distribution representing the distribution of the velocity vectors of moving objects inside the combustion furnace 3 (S103). In S103, the calculation unit 11 performs optical flow estimation based on the acquired video or time-series image. Optical flow estimation is a method for estimating the motion vector of a moving object in an image by analyzing the change in brightness of each pixel contained in the image across multiple images with a time difference. The calculation unit 11 performs optical flow estimation using an existing method. In optical flow estimation, the change in brightness of each pixel is analyzed across multiple frames contained in a video or across multiple still images contained in a time-series image. By optical flow estimation, velocity vectors for each part inside the combustion furnace 3 are obtained.

[0028] In S103, the calculation unit 11 processes the parameters for calculating the velocity vector in regions with a different background color from other regions within the multiple regions obtained by dividing the acquired image, so that they are different from those in other regions. The acquired image shows a heat transfer tube 33. Because the temperature is slightly lower near the heat transfer tube 33, in the infrared image, the region containing the heat transfer tube 33 has a relatively dark background color, while the region not containing the heat transfer tube 33 has a relatively white background color. In other words, the image contains multiple regions with different background colors. In regions with a white background color, the contrast is weak, making it difficult to detect the velocity vector.

[0029] The calculation unit 11 adjusts the processing parameters differently for areas with a white background compared to areas with a black background, so that the processing for calculating the velocity vector is more precise. For example, the calculation unit 11 adjusts the contrast adjustment parameters to enhance the contrast. For example, the calculation unit 11 increases the number of layers in the pyramid image used in optical flow estimation and increases the number of iterations of the calculation for optical flow estimation. Areas with a white background are predetermined, and the calculation unit 11 adjusts the parameters for these predetermined areas. The calculation unit 11 may also determine the background color and identify areas with a relatively white background. By adjusting the parameters in this way, velocity vector detection is reliably performed even in areas with a relatively white background. For areas with a relatively black background, velocity vector detection is sufficiently performed without needing to make the processing for calculating the velocity vector more precise, so no parameter adjustment is performed.

[0030] The velocity vectors are those of moving objects moving inside the combustion furnace 3, and in particular, those of the fluid medium and fuel contained in the bubbling fluidized bed 31. By obtaining the velocity vectors for each part inside the combustion furnace 3, a velocity vector distribution inside the combustion furnace 3 is obtained. The velocity vector distribution includes data indicating the direction and magnitude of the velocity vectors in each part inside the combustion furnace 3. For example, the data indicating the direction and magnitude of the velocity vectors consists of components in three mutually orthogonal directions. The calculation unit 11 generates multiple velocity vector distributions at multiple points in time based on the acquired video or time-series images. The calculation unit 11 stores the generated velocity vector distributions in the storage unit 13.

[0031] The information processing device 1 then outputs heatmaps of the luminance distribution and the velocity vector distribution (S104). In S104, the calculation unit 11 generates a heatmap representing the luminance distribution and a heatmap representing the velocity vector distribution, and displays the heatmaps on the display unit 16. The heatmaps are two-dimensional images that represent the magnitude of the luminance or velocity vector in each part of the interior of the combustion furnace 3 using color. For example, in the luminance distribution heatmap, parts are displayed in red as their luminance increases and in blue as their luminance decreases. For example, in the velocity vector distribution heatmap, parts are displayed in red as their velocity vector is larger and in blue as their velocity vector is smaller. By outputting the heatmaps, the luminance distribution and velocity vector distribution inside the combustion furnace 3 can be visually recognized by the user.

[0032] The calculation unit 11 may display heatmaps of the luminance distribution and the velocity vector distribution simultaneously on the display unit 16, or display them one at a time. The calculation unit 11 may display only one of the heatmaps, either the luminance distribution or the velocity vector distribution. The calculation unit 11 may accept a specification of either the luminance distribution or the velocity vector distribution from the user by operating the operation unit 15, and display the heatmap of the selected distribution.

[0033] The information processing device 1 outputs heatmaps of luminance distribution and velocity vector distribution in a time series (S105). In S105, the calculation unit 11 generates multiple heatmaps of luminance distribution and velocity vector distribution at multiple time points based on the multiple luminance distributions and velocity vector distributions at multiple time points generated. The calculation unit 11 outputs the heatmaps in a time series by displaying the multiple heatmaps of luminance distribution and velocity vector distribution on the display unit 16 in chronological order. For example, the calculation unit 11 displays the multiple heatmaps arranged in chronological order. The calculation unit 11 may also display the multiple heatmaps one by one in sequence. By outputting the heatmaps in a time series, the user can visually recognize the time changes in the luminance distribution and velocity vector distribution inside the combustion furnace 3.

[0034] The calculation unit 11 may display the heatmaps of the luminance distribution and the velocity vector distribution simultaneously in a time series on the display unit 16, or it may display them one at a time. The calculation unit 11 may display only one of the heatmaps of the luminance distribution or the velocity vector distribution in a time series. The calculation unit 11 may accept a specification of either the luminance distribution or the velocity vector distribution from the user by operating the operation unit 15, and display the heatmap of the accepted specification in a time series. The calculation unit 11 may refrain from performing the processing in S105 in response to an instruction received from the user by operating the operation unit 15.

[0035] The information processing device 1 outputs a heatmap of the luminance distribution and a first graph showing the temporal transition of the average luminance (S106). In S106, the calculation unit 11 generates a heatmap of the luminance distribution and a first graph showing the temporal transition of the average luminance based on multiple luminance distributions at multiple time points. At this time, the calculation unit 11 divides the inside of the combustion furnace 3 into multiple areas and generates a heatmap of the luminance distribution and a first graph for each area. The calculation unit 11 displays the generated heatmap of the luminance distribution and the first graph on the display unit 16.

[0036] Figure 5 shows an example of a heatmap of the luminance distribution and the first graph. In Figure 5, the heatmap of the luminance distribution is shown at the top, and the first graph is shown at the bottom. The horizontal axis of the first graph represents time, and the vertical axis represents the average luminance. The horizontal axis of the first graph may also represent elapsed time. In the example shown in Figure 5, the inside of the combustion furnace 3 is divided into four areas, and the temporal transition of the average luminance for each area is displayed in the first graph. That is, the first graph contains multiple graphs showing the temporal transition of the average luminance in each area. The multiple graphs are distinguished from each other by using different display colors. The multiple graphs may also be distinguished by using different display forms of lines or points. The calculation unit 11 calculates the average luminance within each area for each of the multiple luminance distributions, and generates the first graph by connecting the calculated average luminances for each luminance distribution in chronological order. As shown in the example in Figure 5, the calculation unit 11 may calculate the average luminance only for the area corresponding to the bubbling fluidized bed 31 within the area included in the luminance distribution, and may not calculate the average luminance for other areas. The calculation unit 11 may generate and display a first graph of the entire interior of the combustion furnace 3 or the entire bubbling fluidized bed 31.

[0037] The output of the first graph allows the user to easily see the temporal transition of the average brightness inside the combustion furnace 3. By outputting the first graph for each of the multiple areas into which the inside of the combustion furnace 3 is divided, the user can easily see how the partial brightness inside the combustion furnace 3 has changed over time. By outputting a heat map of the brightness distribution and the first graph, the user can compare the heat map and the first graph. The calculation unit 11 overlays an image showing the location of each area onto the heat map. The user can then see the area for which the temporal transition of the average brightness has been calculated.

[0038] The number of areas into which the inside of the combustion furnace 3 is divided may be other than four. For example, the number of areas may be nine or sixteen. The number of areas may be specified by the user by operating the control unit 15. The calculation unit 11 may display the first graph on the display unit 16 and not display the heatmap of the brightness distribution. The calculation unit 11 may receive instructions from the user by operating the control unit 15 to display or hide the heatmap, and may display the heatmap if an instruction to display is received, and not display the heatmap if an instruction to hide is received. The calculation unit 11 may display the heatmap of the brightness distribution in time series. The calculation unit 11 may refrain from performing the processing in S106 in accordance with the instructions received by the user by operating the control unit 15.

[0039] The information processing device 1 outputs a heatmap of the velocity vector distribution and a second graph showing the temporal transition of the average magnitude of the velocity vectors (S107). In S107, the calculation unit 11 generates a heatmap of the velocity vector distribution and a second graph showing the temporal transition of the average magnitude of the velocity vectors based on multiple velocity vector distributions at multiple time points. At this time, the calculation unit 11 generates a heatmap of the velocity vector distribution and a second graph for each of the multiple areas into which the inside of the combustion furnace 3 is divided. The calculation unit 11 displays the generated heatmap of the velocity vector distribution and the second graph on the display unit 16. The calculation unit 11 may further display an infrared image on the display unit 16.

[0040] Figure 6 shows an example of a heatmap and a second graph of the velocity vector distribution. In Figure 6, the infrared image is shown in the upper left, the heatmap of the velocity vector distribution is shown in the upper right, and the second graph is shown at the bottom. The horizontal axis of the second graph represents time, and the vertical axis represents the average magnitude of the velocity vector. The horizontal axis of the second graph may also represent elapsed time. In the example shown in Figure 6, the temporal transition of the average magnitude of the velocity vector for each of the four areas into which the inside of the combustion furnace 3 is divided is displayed in the second graph. That is, the second graph contains multiple graphs showing the temporal transition of the average magnitude of the velocity vector in each area. The multiple graphs are distinguished from each other by using different display colors. The multiple graphs may also be distinguished by using different line or point display formats.

[0041] The calculation unit 11 calculates the average magnitude of the velocity vectors within each area for each of the multiple velocity vector distributions, and generates a second graph by concatenating the calculated average magnitudes of the velocity vectors for each velocity vector distribution in chronological order. As shown in the example in Figure 6, the calculation unit 11 may calculate the average magnitude of the velocity vectors only for the region corresponding to the bubbling fluidized bed 31, and may not calculate the average magnitude of the velocity vectors for other regions. The calculation unit 11 may also generate and display a second graph for the entire interior of the combustion furnace 3 or the entire bubbling fluidized bed 31.

[0042] The output of the second graph allows the user to easily see the temporal transition of the average magnitude of the velocity vector inside the combustion furnace 3. The output of the second graph for each of the multiple areas allows the user to easily understand how the velocity vector has changed in each part of the combustion furnace 3. The output of the velocity vector distribution heatmap and the first graph allows the user to compare the heatmap and the second graph. The calculation unit 11 overlays images showing the location of each area onto the heatmap. The user can then see the area for which the temporal transition of the average magnitude of the velocity vector has been calculated.

[0043] The calculation unit 11 does not have to display the infrared image on the display unit 16. The calculation unit 11 may display the second graph and not display the heatmap of the velocity vector distribution. The calculation unit 11 may receive instructions from the user to display or hide the heatmap by operating the operation unit 15, and may display the heatmap when an instruction to display is received, and not display the heatmap when an instruction to hide is received. The calculation unit 11 may display the heatmap of the velocity vector distribution in time series. The calculation unit 11 may not perform the processing in S107 in response to instructions received from the user by operating the operation unit 15.

[0044] The information processing device 1 outputs either the first or second graph (S108) excluding a specific area from among the multiple areas into which the inside of the combustion furnace 3 is divided. For example, the specific area is the area closest to the fuel inlet 34. The image taken of the area closest to the fuel inlet 34 shows the fuel immediately after it has been put in from the fuel inlet 34. The fuel immediately after being put in has not yet reached the bubbling fluidized bed 31. The brightness distribution or velocity vector distribution for the area closest to the fuel inlet 34 includes the brightness or velocity distribution of the fuel that has not yet reached the bubbling fluidized bed 31, so the state of the bubbling fluidized bed 31 is not adequately reflected. By outputting either the first or second graph excluding the area closest to the fuel inlet 34, the first or second graph that reflects the state of the bubbling fluidized bed 31 is output.

[0045] In S108, the calculation unit 11 generates a luminance distribution or velocity vector distribution excluding a specific area based on the luminance distribution or velocity vector distribution, and generates a first graph or second graph excluding the specific area based on the generated luminance distribution or velocity vector distribution. The calculation unit 11 may also generate a first graph or second graph excluding a specific area by excluding the first graph or second graph for a specific area from the first graph or second graph for a plurality of areas that have already been generated. The calculation unit 11 may also generate a first graph or second graph for the entire plurality of areas excluding the specific area.

[0046] The calculation unit 11 displays the generated first or second graph on the display unit 16. For example, information identifying the area closest to the fuel inlet 34 is pre-stored in the storage unit 13, and the calculation unit 11 identifies the area based on the stored information and generates a brightness distribution or velocity vector distribution excluding the identified area. The calculation unit 11 may also generate a brightness distribution or velocity vector distribution excluding a specific area from an infrared image. By outputting the first or second graph excluding the area closest to the fuel inlet 34, the user can easily confirm the temporal transition of the average brightness or average velocity vector magnitude in the bubbling fluidized bed 31. Therefore, the user can observe the bubbling fluidized bed 31 in detail.

[0047] The calculation unit 11 may display both the first graph and the second graph on the display unit 16, or it may display either the first graph or the second graph. The calculation unit 11 may receive an instruction from the user to specify the first graph or the second graph by operating the operation unit 15, and may display the specified graph. The calculation unit 11 may generate and display a heatmap of the luminance distribution or velocity vector distribution excluding a specific area. The calculation unit 11 may receive an instruction from the user to display or hide the heatmap by operating the operation unit 15, and may display or hide the heatmap. The calculation unit 11 may refrain from performing the processing in S108 in response to an instruction received from the user by operating the operation unit 15.

[0048] The information processing device 1 outputs a heatmap and a second graph of the component of the velocity vector distribution in a specific direction (S109). For example, the specific direction is the horizontal or vertical direction. The horizontal or vertical component of the velocity vector distribution indicates the velocity distribution of the fuel or fluid medium contained in the bubbling fluidized bed 31 as it moves horizontally or vertically. In S109, the calculation unit 11 extracts the component in the specific direction from the velocity vector distribution, generates a heatmap and a second graph based on the extracted component, and displays the generated heatmap and second graph on the display unit 16. The calculation unit 11 may generate and display a second graph for each of the multiple areas into which the inside of the combustion furnace 3 is divided, or it may generate and display a second graph for the entire inside of the combustion furnace 3 or for the entire bubbling fluidized bed 31.

[0049] By outputting a heatmap or second graph of the component of the velocity vector distribution in a specific direction, the user can easily confirm the velocity distribution of a moving object in a specific direction, or the temporal transition of the average velocity in a specific direction. By outputting a heatmap or second graph in the horizontal or vertical direction, the user can easily confirm the velocity distribution or the temporal transition of the average velocity in the horizontal or vertical direction. More specifically, the user can easily confirm how the fuel or fluid medium contained in the bubbling fluidized bed 31 moves in the horizontal or vertical direction.

[0050] The calculation unit 11 does not have to display the heatmap on the display unit 16. The calculation unit 11 may receive an instruction from the user to display or hide the heatmap by operating the operation unit 15, and may display or hide the heatmap. The calculation unit 11 may display a heatmap of the component of the velocity vector distribution in a specific direction in time series. The specific direction may be specified by the user by operating the operation unit 15. The specific direction may be specified by area. The calculation unit 11 displays the heatmap and the second graph of the component in the specified direction. The calculation unit 11 may refrain from performing the processing in S109 in response to an instruction received from the user by operating the operation unit 15.

[0051] The information processing device 1 determines whether or not to terminate the process for observing the inside of the combustion furnace 3 (S110). If the user operates the operation unit 15 and receives an instruction to terminate the process, the calculation unit 11 determines to terminate the process. If no instruction to terminate the process is received, the calculation unit 11 determines not to terminate the process. If it is determined not to terminate the process (S110: NO), the information processing device 1 returns to process S104. If it is determined to terminate the process (S110: YES), the information processing device 1 terminates the process for observing the inside of the combustion furnace 3. The information processing device 1 executes processes S101 to S110 as needed.

[0052] The information processing device 1 performs information processing to detect anomalies inside the combustion furnace 3 based on the brightness distribution and the velocity vector distribution. The information processing device 1 includes an anomaly detection model 132 that outputs an anomaly score indicating the degree of anomaly inside the combustion furnace 3 when the features of the brightness distribution and the velocity vector distribution are input. The information processing device 1 also includes a first feature extraction model 133 that extracts feature quantities of the brightness distribution and a second feature extraction model 134 that extracts feature quantities of the velocity vector distribution.

[0053] Figure 7 is a conceptual diagram showing examples of the functions of the anomaly detection model 132, the first feature extraction model 133, and the second feature extraction model 134. The first feature extraction model 133 receives a luminance distribution as input, and the second feature extraction model 134 receives a velocity vector distribution as input. The first feature extraction model 133 is a pre-trained model that has been pre-trained to output the features of a luminance distribution when a luminance distribution is input. The second feature extraction model 134 is a pre-trained model that has been pre-trained to output the features of a velocity vector distribution when a velocity vector distribution is input. For example, the first feature extraction model 133 and the second feature extraction model 134 are encoders implemented using a CNN (Convolutional Neural Network). For example, the first feature extraction model 133 can be created by training an autoencoder to output a luminance distribution that matches the input luminance distribution, and then using the encoder portion of the trained autoencoder as the first feature extraction model 133. The second feature extraction model 134 may be created in a similar manner.

[0054] The anomaly detection model 132 is a pre-trained model that, when given luminance distribution features and velocity vector distribution features, outputs an anomaly score indicating the degree of anomaly inside the combustion furnace 3. The anomaly score is numerical information indicating the degree of anomaly inside the combustion furnace 3. For example, the higher the degree of anomaly, the larger the anomaly score. For example, the anomaly score can take any number between 0 and 100. The anomaly score may also be information indicating the degree of anomaly in a form other than a numerical value. For example, the anomaly score may be information indicating one of multiple levels such as high anomaly, medium anomaly, and low anomaly.

[0055] The anomaly detection model 132 is implemented using, for example, a Oneclass SVM or a neural network. The anomaly detection model 132 is trained using training data that includes acquired luminance distribution features and velocity vector distribution features. For example, the training data includes data in which luminance distribution features and velocity vector distribution features are associated with an anomaly score determined by a human. For example, the training data consists of data in which luminance distribution features and velocity vector distribution features obtained for a normal combustion furnace 3 are associated with an anomaly score indicating a low degree of anomaly. The anomaly detection model 132 is trained to output an anomaly score indicating a high degree of anomaly when features different from those of the luminance distribution features and velocity vector distribution features in a normal state are input.

[0056] The first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132 are realized by the arithmetic unit 11 executing information processing according to the computer program 131. The storage unit 13 stores the data necessary to realize the first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132.

[0057] The first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132 may be configured using hardware. For example, the first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132 may be configured using hardware including a processor and memory for storing the necessary programs and data. Alternatively, the first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132 may be implemented using a quantum computer. Alternatively, the first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132 may be provided outside the information processing device 1, and the information processing device 1 may execute processing using the external first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132. For example, the first feature extraction model 133, the second feature extraction model 134, and the anomaly detection model 132 may be implemented using multiple computers connected via a communication network, or they may be implemented using a cloud.

[0058] Figure 8 is a flowchart illustrating an example of the procedure performed by the information processing device 1 to detect an anomaly inside the combustion furnace 3. The information processing device 1 generates feature quantities for the luminance distribution and the velocity vector distribution (S21). In S21, the calculation unit 11 inputs the luminance distribution to the first feature extraction model 133. The first feature extraction model 133 performs calculations in response to the input of the luminance distribution and outputs the feature quantities for the luminance distribution. The calculation unit 11 acquires the feature quantities for the luminance distribution output by the first feature extraction model 133. The calculation unit 11 also inputs the velocity vector distribution to the second feature extraction model 134. The second feature extraction model 134 performs calculations in response to the input of the velocity vector distribution and outputs the feature quantities for the velocity vector distribution. The calculation unit 11 acquires the feature quantities for the velocity vector distribution output by the second feature extraction model 134.

[0059] The information processing device 1 inputs the feature quantities of the luminance distribution and the velocity vector distribution to the anomaly detection model 132 (S22). In S22, the calculation unit 11 inputs the acquired feature quantities of the luminance distribution and the velocity vector distribution to the anomaly detection model 132. The anomaly detection model 132 performs calculations according to the input of the feature quantities of the luminance distribution and the velocity vector distribution and outputs an anomaly score. The information processing device 1 acquires the anomaly score (S23). In S23, the calculation unit 11 acquires the anomaly score output by the anomaly detection model 132.

[0060] The information processing device 1 outputs the degree of abnormality inside the combustion furnace 3 (S24). In S24, the calculation unit 11 determines whether or not the inside of the combustion furnace 3 is abnormal based on the acquired degree of abnormality score, and if the inside of the combustion furnace 3 is abnormal, it displays on the display unit 16 that the inside of the combustion furnace 3 is abnormal. For example, the calculation unit 11 determines that the inside of the combustion furnace 3 is abnormal if the degree of abnormality score exceeds a predetermined threshold. The calculation unit 11 may also display on the display unit 16 that the inside of the combustion furnace 3 is not abnormal if it is not abnormal. The calculation unit 11 may display the value of the degree of abnormality score on the display unit 16, display the degree of abnormality inside the combustion furnace 3 according to the degree of abnormality score, or display a warning on the display unit 16 according to the degree of abnormality score.

[0061] The information processing device 1 outputs the degree of abnormality inside the combustion furnace 3, allowing the user to easily determine whether the inside of the combustion furnace 3 is abnormal or not. For example, if a flow defect occurs in the bubbling fluidized bed 31, such as when the magnitude of the velocity vector becomes extremely small in a specific part of the bubbling fluidized bed 31, a high degree of abnormality will be output, informing the user that the inside of the combustion furnace 3 is abnormal. The user can then appropriately manage the combustion furnace 3, such as by inspecting it according to the abnormality.

[0062] After S24 is completed, the information processing device 1 terminates the process for detecting abnormalities inside the combustion furnace 3. The information processing device 1 executes the processes from S21 to S24 as needed. The first feature extraction model 133 may be configured to output a first feature quantity when a heatmap of luminance distribution is input, and the second feature extraction model 134 may be configured to output a second feature quantity when a heatmap of velocity vector distribution is input. In this configuration, the calculation unit 11 inputs the heatmap of luminance distribution to the first feature extraction model 133 and the heatmap of velocity vector distribution to the second feature extraction model 134 in S21, and obtains the first feature quantity output by the first feature extraction model 133 and the second feature quantity output by the second feature extraction model 134. In S22, the calculation unit 11 inputs the first feature quantity and the second feature quantity to the anomaly detection model 132, obtains the anomaly score output by the anomaly detection model 132 in S23, and outputs the anomaly score according to the anomaly score in S24.

[0063] The first feature extraction model 133 may output a first feature when a luminance distribution or a heatmap of a luminance distribution is input in a time series. The second feature extraction model 134 may output a second feature when a velocity vector distribution or a heatmap of a velocity vector distribution is input in a time series. The luminance distribution feature or the velocity vector distribution feature may be extracted by a method other than using a trained model.

[0064] The anomaly detection model 132, the first feature extraction model 133, and the second feature extraction model 134 may be implemented as a single pre-trained model. This pre-trained model is pre-trained to output an anomaly score when a luminance distribution and a velocity vector distribution are input. The pre-trained model is trained using training data that includes the obtained luminance distribution and velocity vector distribution and a defined anomaly score. The information processing device 1 inputs the luminance distribution and velocity vector distribution to the pre-trained model, performs calculations using the pre-trained model, and obtains the anomaly score output by the pre-trained model, thereby performing the processing corresponding to S21 to S23. The pre-trained model may also output an anomaly score when a heatmap of the luminance distribution and a heatmap of the velocity vector distribution are input.

[0065] The information processing device 1 may be configured to control other devices included in the plant 2 and related to the combustion furnace 3, depending on the degree of abnormality inside the combustion furnace 3. Other devices related to the combustion furnace 3 include, for example, a ventilator, a fluidized medium discharger, a fluidized medium sorting and processing device, a fluidized medium circulation device, or a fluidized medium replenishment device. In this configuration, the information processing device 1 is connected to the other devices related to the combustion furnace 3. Figure 9 is a flowchart showing an example of a procedure for controlling other devices in response to an abnormality inside the combustion furnace 3. The information processing device 1 performs the same processing as S21 to S24, from S31 to S34. After S34 is completed, the information processing device 1 controls the other devices related to the combustion furnace 3 according to the degree of abnormality inside the combustion furnace 3 (S35). The information processing device 1 may skip processing S34 and proceed to processing S35 after S33 is completed.

[0066] In S35, the calculation unit 11 controls other devices by sending control signals from the input / output unit 17 to them according to the degree of abnormality inside the combustion furnace 3. The other devices operate according to the control signals sent from the information processing device 1. For example, the calculation unit 11 sends a control signal when the degree of abnormality reaches a predetermined threshold. The information processing device 1, for example, causes the ventilation device to change the airflow rate when the degree of abnormality reaches a predetermined threshold. The information processing device 1, for example, causes the fluidized medium circulation device to clean the fluidized medium when the degree of abnormality reaches a predetermined threshold. The information processing device 1, for example, causes the fluidized medium discharge device to discharge at least a portion of the fluidized medium and the fluidized medium replenishment device to replenish the fluidized medium, thereby replacing the fluidized medium when the degree of abnormality reaches a predetermined threshold. After S35 is completed, the information processing device 1 terminates the process of controlling other devices according to the abnormality inside the combustion furnace 3. The information processing device 1 executes the processes of S31 to S35 as needed. The information processing device 1 may execute the processes of S31 to S35 in parallel with the processes of S21 to S24.

[0067] The processing in S31 to S35 causes other devices related to the combustion furnace 3 to operate according to the degree of abnormality inside the combustion furnace 3, and corrective actions are taken to improve the state of the bubbling fluidized bed 31. For example, by taking corrective actions when the degree of abnormality reaches a predetermined threshold, corrective actions can be taken before the flow state of the bubbling fluidized bed 31 deteriorates. By taking corrective actions before the flow state deteriorates, it is possible to maintain the flow state appropriately and consume the fluidized medium without excess or deficiency. The information processing device 1 may also perform processing to determine the degree of abnormality inside the combustion furnace 3 without using a trained model. For example, the information processing device 1 may perform processing to determine the degree of abnormality based on the magnitude of the difference in brightness inside the combustion furnace 3.

[0068] As described in detail above, in this embodiment, the information processing device 1 acquires images of the inside of the combustion furnace 3 taken by the infrared camera 32, generates a brightness distribution and a velocity vector distribution inside the combustion furnace 3, and outputs analysis results based on the brightness distribution and velocity vector distribution. For example, the information processing device 1 outputs a heat map and a first graph of the brightness distribution, and a heat map and a second graph of the velocity vector distribution. By checking the outputted analysis results, the user can observe the inside of the combustion furnace 3 in detail. For example, the user can observe the state of the bubbling fluidized bed 31 inside the combustion furnace 3. Furthermore, by controlling other devices related to the combustion furnace 3 according to the observed state inside the combustion furnace 3, it is possible to maintain an appropriate flow state of the bubbling fluidized bed 31.

[0069] In this embodiment, the information processing device 1 outputs information by displaying it on the display unit 16, but the information processing device 1 may output information by other means. The information processing system 100 includes a terminal device that can communicate with the information processing device 1, and the information processing device 1 may transmit information to the terminal device, and the terminal device may display the information on its display unit. For example, analysis results based on the luminance distribution and velocity vector distribution may be output by displaying a heat map of the luminance distribution and velocity vector distribution, as well as a first graph and a second graph, on the display unit of the terminal device.

[0070] The present invention is not limited to the embodiments described above, and various modifications are possible within the scope of the claims. That is, embodiments obtained by combining technical means that have been appropriately modified within the scope of the claims are also included in the technical scope of the present invention.

[0071] The matters described in each embodiment can be combined with each other. Furthermore, the independent and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. Moreover, although the claims do not use a form in which claims referencing two or more other claims (multi-claim form), they are not limited to this. They may be described using a multi-claim form, or a form in which multi-claims referencing at least one multi-claim (multi-multi-claim). [Explanation of Symbols]

[0072] 100 Information Processing Systems 1. Information Processing Device 10 Recording media 11 Arithmetic section 13 Storage section 131 Computer Programs 132 Anomaly Detection Models 2 Plants 3. Combustion furnace (bubbling fluidized bed combustion furnace) 31 Bubbling fluidized bed 32 Infrared Cameras 33 Heat transfer tubes 34 Fuel inlet

Claims

1. Images were taken of the inside of a bubbling fluidized bed combustion furnace using an infrared camera. Based on the aforementioned image, a luminance distribution inside the bubbling fluidized bed combustion furnace and a velocity vector distribution representing the distribution of the velocity vectors of moving objects inside the bubbling fluidized bed combustion furnace are generated. Outputs analysis results based on the luminance distribution and the velocity vector distribution. A computer program characterized by causing a computer to perform a process.

2. Output heatmaps of the luminance distribution and velocity vector distribution at multiple points in time in a time series. The computer program according to claim 1, characterized in that it causes a computer to perform a process.

3. Based on the luminance distribution at multiple points in time, a heatmap of the luminance distribution and a first graph showing the temporal transition of the average luminance are output. Based on the velocity vector distribution at multiple time points, a heatmap of the velocity vector distribution and a second graph showing the temporal transition of the average magnitude of the velocity vectors are output. The computer program according to claim 1, characterized in that it causes a computer to perform a process.

4. For each of the multiple areas into which the interior of the bubbling fluidized bed combustion furnace is divided, the first graph or the second graph is output. The computer program according to claim 3, characterized in that it causes a computer to perform the processing.

5. The first graph or the second graph is output for the area excluding a specific area from among the multiple areas into which the interior of the bubbling fluidized bed combustion furnace is divided. The computer program according to claim 3, characterized in that it causes a computer to perform the processing.

6. The output displays a heatmap of the component of the velocity vector distribution in a specific direction, and a second graph showing the temporal transition of the average magnitude of the component. The computer program according to claim 3, characterized in that it causes a computer to perform the processing.

7. For the region in the divided image where the background color is different from the other regions, the parameters for calculating the velocity vector are made different from those for the other regions. The computer program according to claim 1, characterized in that it causes a computer to perform a process.

8. Using a trained model that outputs a score indicating the degree of abnormality inside the bubbling fluidized bed combustion furnace according to the brightness distribution and velocity vector distribution, the score corresponding to the acquired brightness distribution and velocity vector distribution is obtained. The computer program according to claim 1, characterized in that it causes a computer to perform a process.

9. The first feature of the heatmap of the luminance distribution and the second feature of the heatmap of the velocity vector distribution are obtained. The acquired first and second features are input to the trained model that outputs the score when the first and second features are input, and the score obtained by the trained model is retrieved. Based on the acquired score, the degree of abnormality inside the bubbling fluidized bed combustion furnace is output. The computer program according to claim 8, characterized in that it causes a computer to perform a process.

10. Depending on the degree of abnormality inside the bubbling fluidized bed combustion furnace, control other devices associated with the bubbling fluidized bed combustion furnace. The computer program according to claim 9, characterized in that it causes a computer to perform a process.

11. Images were taken of the inside of a bubbling fluidized bed combustion furnace using an infrared camera. Based on the aforementioned image, a luminance distribution inside the bubbling fluidized bed combustion furnace and a velocity vector distribution representing the distribution of the velocity vectors of moving objects inside the bubbling fluidized bed combustion furnace are generated. Outputs analysis results based on the luminance distribution and the velocity vector distribution. An information processing method characterized by the following:

12. Equipped with a calculation unit, The aforementioned arithmetic unit, Images were taken of the inside of a bubbling fluidized bed combustion furnace using an infrared camera. Based on the aforementioned image, a luminance distribution inside the bubbling fluidized bed combustion furnace and a velocity vector distribution representing the distribution of the velocity vectors of moving objects inside the bubbling fluidized bed combustion furnace are generated. Outputs analysis results based on the luminance distribution and the velocity vector distribution. An information processing device characterized by the following:

Images