Biomass fuel inappropriate material detection system, biomass fuel inappropriate material detection program, biomass fuel inappropriate material detection method, and installation method

Abstract

To provide a biomass fuel inappropriate material detection system, a biomass fuel inappropriate material detection program, and a biomass fuel inappropriate material detection method which detect a fuel inappropriate material from a biomass fuel, and a method of installing an apparatus of removing a fuel inappropriate material.SOLUTION: A biomass fuel inappropriate material detection system includes a sensor which detects a mixture including a biomass fuel, and a control part. The control part detects a fuel inappropriate material in the mixture on the basis of sensor data acquired by the sensor.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a biomass fuel unsuitable substance detection system for detecting unsuitable substances for fuel from biomass fuel, a biomass fuel unsuitable substance detection program, a biomass fuel unsuitable substance detection method, and a method for installing an unsuitable substance for fuel removal device. [Background technology]

[0002] In recent years, biomass power generation has been attracting attention. Biomass power generation refers to generating electricity using renewable biological resources (biomass) such as wood and plant residues as raw materials. When biomass is burned, carbon dioxide is produced just like fossil fuels, but because plants absorb the carbon dioxide and grow, reproducing biomass, overall it can be considered that the amount of carbon dioxide in the atmosphere does not increase. Biomass power generation is a carbon-neutral method of generating electricity, and the electricity it produces is renewable energy, so it is attracting attention.

[0003] Biomass power generation includes a direct fuel system that uses steam obtained by directly burning biomass as fuel to generate power. Examples of biomass fuel (hereinafter referred to as "biomass fuel") include wood chips, bark, wood pellets, PKS (Palm Kernel Shell), livestock manure, rice husk, and corn meal.

[0004] Biomass fuels accepted by biomass power plants that generate biomass power are a mixture of various raw materials, and are not homogenous in terms of type and shape. Therefore, the accepted mixture may contain unsuitable fuels, such as biomass fuels of inappropriate size or metals that are not biomass fuels. Unsuitable fuels can cause problems such as blockages and jamming in fuel transport equipment such as conveyors and chutes. If a problem causes the fuel transport to become unstable, the amount of power generated will decrease, and revenue from selling electricity will decrease. However, unsuitable fuels are not constantly monitored, and if a problem occurs, the inclusion of unsuitable fuels will become apparent.

[0005] A similar technology to the technology for detecting unsuitable objects is a technology for detecting objects unsuitable for combustion at waste incineration facilities. Patent Document 1 proposes an unsuitable object detection system that photographs waste brought into a waste incineration facility and detects objects unsuitable for incineration from the captured images of the waste. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2019-190805 A Summary of the Invention [Problem to be solved by the invention]

[0007] However, the conventional technology is for detecting materials unsuitable for incineration from garbage, and cannot be applied to detecting materials unsuitable for combustion from mixtures brought in as biomass fuel. The present invention has been made in consideration of such circumstances. Its purpose is to provide a system for detecting materials unsuitable for incineration from a mixture containing biomass fuel, a program for detecting materials unsuitable for incineration from a mixture containing biomass fuel, a method for detecting materials unsuitable for incineration from a mixture containing biomass fuel, and a method for installing a removal device for materials unsuitable for incineration from a mixture containing biomass fuel. [Means for solving the problem]

[0008] A biomass fuel unsuitable substance detection system according to one embodiment of the present application includes a sensor that detects a mixture containing biomass fuel, and a control unit, and the control unit detects fuel unsuitable substances in the mixture based on sensor data acquired by the sensor.

[0009] In one embodiment of the present application, it becomes possible to detect fuel unsuitable substances from a mixture containing biomass fuel.

[0010] In a biomass fuel unsuitable substance detection system according to one embodiment of the present application, the sensor is a camera that photographs the mixture, and the control unit acquires mixture image data including the mixture, and when the mixture image data is input, outputs a detection result of the unsuitable substances for fuel by inputting the acquired mixture image data into a learning model that has been trained to recognize unsuitable substances for fuel, including long objects and metals.

[0011] In one aspect of the present application, the sensor is a camera that photographs the mixture, and image data of the mixture photographed by the camera is input into a learning model that has been trained to recognize unsuitable fuel substances, including long objects and metals, and the unsuitable fuel substances are detected, making it possible to detect the unsuitable fuel substances with high accuracy.

[0012] In one embodiment of the biomass fuel unsuitable substance detection system of the present application, the camera is installed at a monitoring port of a quantitative feeder.

[0013] In one aspect of the present application, since uncombustible materials are detected in a quantitative feeder, it is possible to reduce the range of problems caused by uncombustible materials.

[0014] In a biomass fuel unsuitable substance detection system according to one aspect of the present application, the number of detected unsuitable substances for fuel is counted for each type.

[0015] In one aspect of the present application, the number of unsuitable fuel materials is counted by type, making it possible to grasp the mixing status of unsuitable fuel materials.

[0016] In a biomass fuel unsuitable substance detection system according to one aspect of the present application, an unsuitable substance image including the detected unsuitable substance for fuel is output.

[0017] In one aspect of the present application, an image of the unsuitable object is output, so that the unsuitable object for combustion can be visually confirmed.

[0018] In a biomass fuel unsuitable substance detection system according to one aspect of the present application, the unsuitable substance image has a graphic depiction indicating an area of ​​the unsuitable substance for fuel.

[0019] In one aspect of the present application, a figure showing the area of ​​the fuel unsuitable object is drawn, so that the location of the fuel unsuitable object in the image can be easily confirmed.

[0020] In the biomass fuel unsuitable substance detection system according to one aspect of the present application, the graphic is displayed in a different manner depending on attributes including the type of the unsuitable substance for fuel.

[0021] In one aspect of the present application, the display form of the graphic is made different depending on the attribute including the type of the unsuitable fuel substance, so that the attribute of the unsuitable fuel substance such as the type can be confirmed.

[0022] In one embodiment of the biomass fuel unsuitable materials detection system of the present application, the sensor is a microphone, and the control unit acquires mixture voice data generated when the mixture moves, and when the mixture voice data is input, outputs the presence or absence of unsuitable fuel materials by inputting the acquired mixture voice data into a second learning model that is trained to recognize unsuitable fuel materials including long objects and metals.

[0023] In one aspect of the present application, the sensor is a microphone and collects sound generated when the mixture moves. The collected mixture sound data is input into a learning model trained to recognize unsuitable fuels, including long objects and metals, and unsuitable fuels are detected, thereby making it possible to detect unsuitable fuels with high accuracy.

[0024] In a biomass fuel unsuitable substance detection system according to one aspect of the present application, the microphone is installed in a chute that forms a path through which the mixture falls.

[0025] In one aspect of the present application, a microphone is installed in the chute that constitutes the path along which the mixture falls, making it possible to carry out procedures for discharging the unsuitable for combustion immediately after it has been carried out of the chute.

[0026] In a biomass fuel unsuitable material detection system according to one embodiment of the present application, when the unsuitable fuel is detected inside the chute, the unsuitable fuel is re-recognized on the conveyor on which it is being transported after being discharged from the chute, and the unsuitable fuel is discharged by switching the damper of the conveyor, or the biofuel supply port is closed to discharge the unsuitable fuel together with the surplus biofuel.

[0027] In one embodiment of the present application, it is possible to discharge materials unsuitable for combustion.

[0028] In a biomass fuel unsuitable substance detection system according to one aspect of the present application, audio data of the unsuitable substance for fuel is output together with image data.

[0029] In one aspect of the present application, audio data of the unsuitable fuel object is output together with image data, so that it is possible to estimate the weight of the detected unsuitable fuel object.

[0030] A biomass fuel unsuitable substance detection program according to one embodiment of the present application detects a mixture containing biomass fuel using a sensor, and detects fuel unsuitable substances in the mixture based on sensor data acquired by the sensor.

[0031] A method for detecting unsuitable substances for biomass fuel according to one aspect of the present application detects a mixture containing biomass fuel with a sensor, and detects unsuitable substances for fuel in the mixture based on sensor data acquired by the sensor.

[0032] In one embodiment of the present application, it becomes possible to detect fuel unsuitable substances from a mixture containing biomass fuel.

[0033] An installation method according to one aspect of the present application involves obtaining mixture image data containing biomass fuel using a camera attached to a biomass fuel processing facility at a biomass fuel power plant where no unsuitable fuel removal equipment is installed, inputting the obtained mixture image data into a learning model that has been trained to recognize unsuitable fuels including long objects and metals when the mixture image data is input, thereby outputting the unsuitable fuels, outputting the type and number of the unsuitable fuels that have been output, and installing the unsuitable fuel removal equipment based on the mounting position of the camera.

[0034] In one aspect of the present application, a device for removing unsuitable fuel materials is installed based on the mounting position of the camera, making it possible to appropriately remove detected unsuitable fuel materials from a mixture containing biomass fuel. Effect of the Invention

[0035] In one aspect of the present application, it becomes possible to detect fuel unsuitable substances from a mixture containing biomass fuel. [Brief description of the drawings]

[0036] [Figure 1] FIG. 2 is an explanatory diagram showing a configuration example of a fuel supply system. [Diagram 2] FIG. 2 is a block diagram showing an example of a hardware configuration of a control device. [Diagram 3] FIG. 11 is an explanatory diagram illustrating an example of a discharge mechanism. [Figure 4] FIG. 11 is an explanatory diagram showing an example of an image DB. [Diagram 5] FIG. 13 is an explanatory diagram showing an example of a training image DB. [Figure 6] FIG. 2 is an explanatory diagram showing an example of a voice DB. [Figure 7] FIG. 2 is an explanatory diagram showing an example of a training voice DB. [Figure 8] FIG. 2 is an explanatory diagram showing an example of a first learning model. [Figure 9] 13 is a flowchart illustrating an example of a procedure for a first generation process. [Figure 10] FIG. 11 is an explanatory diagram showing an example of a second learning model. [Figure 11] 13 is a flowchart illustrating an example of a procedure for a second generation process. [Figure 12] 13 is a flowchart illustrating an example of a procedure for a determination process. [Figure 13] 13 is a flowchart illustrating an example of a procedure for a counting process. [Figure 14] FIG. 13 is an explanatory diagram showing an example of a tally result display screen. [Figure 15] 13 is a flowchart showing an example of a procedure for a confirmation screen display process. [Figure 16] FIG. 13 is an explanatory diagram showing an example of a list screen. [Figure 17] FIG. 13 is an explanatory diagram showing an example of a confirmation screen. [Figure 18] 1 is a flowchart showing the flow of an implementation support operation. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0037] Hereinafter, an embodiment will be described with reference to the drawings. FIG. 1 is an explanatory diagram showing an example of the configuration of a fuel supply system. The fuel supply system 100 constitutes a part of a biomass fuel power plant. The fuel supply system 100 supplies biomass fuel (biofuel) to a boiler combustion chamber 80 that generates steam for turning a steam turbine connected to a generator. The fuel supply system 100 includes a fuel receiving section 10, a supply conveying section 20, a pushing conveying section 30, a fuel input section 40, a circulating conveying section 50, a control device 60, and a monitoring section 70. Note that for convenience of drawing, the monitoring section 70 is not shown in FIG. 1, but the components of the monitoring section 70 (such as a camera 71) described later are shown.

[0038] The fuel receiving section 10 is, for example, a facility with an open top. The fuel receiving section 10 stores biomass fuel. The biomass fuel is dropped into the fuel receiving section 10 by a wheel loader or the like. The fuel receiving section 10 includes a leveling conveyor 11, a picker roller 12, a fixed-volume feeder 13, and a damper 14.

[0039] The scraping / leveling conveyor 11 regulates the width and height of the biomass fuel input from the circulating transport section 50. The picker rollers 12 pick up a predetermined amount of the biomass fuel on the scraping / leveling conveyor 11 and transport it to the supply transport section 20.

[0040] The quantitative feeder 13 conveys a predetermined amount of the biomass fuel dropped by a wheel loader or the like to the supply and conveyance section 20. If it is determined that the biomass fuel conveyed by the quantitative feeder 13 contains unsuitable fuel substances, the damper 14 discharges the biomass fuel (mixture) containing the unsuitable fuel substances to the outside so that it is not conveyed to the supply and conveyance section 20.

[0041] The supply conveying section 20 conveys the biomass fuel discharged from the fuel receiving section 10 to the push-in conveying section 30. The supply conveying section 20 is a hollow conveying path having an enclosure 21. The biomass fuel is transported by a transport mechanism provided inside the enclosure 21. For example, the supply conveying section 20 is composed of a flight conveyor. The supply conveying section 20 transports the biomass fuel discharged from the fuel receiving section 10 located below diagonally upward. The transported biomass fuel is transported from the upper end 22. The biomass fuel transported from the upper end 22 is supplied to the push-in conveying section 30 via a falling path including a chute 23.

[0042] The push-in conveying section 30 pushes and supplies the supplied biomass fuel to the fuel input section 40. The push-in conveying section 30 includes a flight conveyor 31, a slide gate 32, and a discharge port 33. The flight conveyor 31 has a plurality of rectangular plate-shaped blade members spaced apart on its conveying surface.

[0043] If it is determined that the biomass fuel passing through the chute 23 contains unsuitable fuel substances, the slide gate 32 is closed and the biomass fuel containing the unsuitable fuel substances is discharged from the discharge port 33. If it is determined that the biomass fuel does not contain unsuitable fuel substances, the slide gate 32 is left open, and the surplus fuel that could not be contained in the fuel input section 40 is carried out to the circulating transport section 50.

[0044] The fuel input section 40 introduces biomass fuel into the boiler combustion chamber 80. The fuel input section 40 includes a slide gate 41, a chute 42, a hopper 43, and a supply device 44. In FIG. 1, two sets of slide gates 41, chutes 42, hoppers 43, and supply devices 44 are illustrated. When it is determined that the biomass fuel passing through the chutes 23 contains unsuitable fuels, each slide gate 41 (biofuel supply port) is closed to prevent the unsuitable fuels from being sent to the boiler combustion chamber 80 of the boiler. When it is determined that the biomass fuel does not contain unsuitable fuels, each slide gate 41 is opened. The biomass fuel falls from each chute 42 through the slide gate 41 by the blade member moving on the underside of the flight conveyor 31 described above. The biomass fuel is supplied to each hopper 43 through each chute 42. The biomass fuel charged and filled in each hopper 43 is sent to the boiler combustion chamber 80 while the supply amount is adjusted by the supply device 44 provided below each hopper.

[0045] In the fuel input section 40, if it is determined that the biomass fuel passing through the chute 23 contains unsuitable fuel, the slide gate 41 is closed to prevent the unsuitable fuel from being accepted. At this time, if the time elapsed from the position where the unsuitable fuel is detected until it reaches the slide gate 41 is not accurate, the unsuitable fuel may pass through the slide gate 41. To prevent this, a camera may be installed in the push-in transport section 30 to re-identify the unsuitable fuel and control the opening and closing of the slide gate 41.

[0046] The circulating transport section 50 returns the surplus fuel that could not be accommodated in the fuel input section 40 from the pushing transport section 30 to the fuel receiving section 10. Even if the fuel unsuitable substances are not detected until they reach the circulating transport section 50, they will be detected again after being returned to the fuel receiving section 10, so there is a good chance that they can be discharged.

[0047] The control device 60 controls the fuel supply system 100. The control device 60 is configured with an industrial PC (Personal Computer), a panel computer, a server computer, or the like.

[0048] The monitoring unit 70 includes three cameras 71, 72, and 73 and a microphone 74. The camera 71 is installed at the monitoring port of the quantitative feeder 13. The camera 71 photographs the biomass fuel being conveyed from the fuel receiving unit 10 to the supplying and conveying unit 20. The camera 71 is installed at the monitoring port of the chute 23. The camera 72 photographs the biomass fuel being fed into the chute 23 from the upper end 22 of the supplying and conveying unit 20. The camera 73 is installed at the monitoring port of the circulating and conveying unit 50. The camera 73 photographs the biomass fuel being returned from the circulating and conveying unit 50 to the fuel receiving unit 10. The microphone 74 is installed at the chute 23. The microphone 74 is installed for the purpose of collecting sounds generated when the biomass fuel is fed into the chute 23 and passes through the chute 23. Images captured by each camera are sent to the control device 60. The sounds collected by the microphone 74 are A / D converted by an A / D converter (not shown) or the like, and sent to the control device 60.

[0049] 2 is a block diagram showing an example of the hardware configuration of the control device. The control device 60 includes a control unit 61, an input / output unit 62, a storage unit 63, a communication unit 64, a display unit 65, and an operation unit 66. The control unit 61 includes an arithmetic processing device such as a CPU (Central Processing Unit), an MPU (Micro-Processing Unit) or a GPU (Graphics Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory). The control unit 61 reads out and executes a program 60P stored in the storage unit 63 to perform various processes.

[0050] The input / output unit 62 inputs and outputs data between the cameras 71, 72, and 73. The input / output unit 62 is connected to the cameras 71, 72, and 73 via, for example, a signal line, and inputs and outputs data by serial communication or parallel communication via the signal line. The input / output unit 62 transmits data such as control commands provided by the control unit 61 to the cameras 71, 72, and 73, and also provides image data input from the cameras 71, 72, and 73 to the control unit 61. Note that communication between the cameras 71, 72, and 73 and the control device 60 may be short-distance wireless communication, for example, Bluetooth (registered trademark), an ad-hoc mode of a wireless LAN, or ZigBee (registered trademark), or communication using a LAN. The input / output unit 62 also accepts voice data obtained by digitizing voice collected by the microphone 74. Communication between the input / output unit 62 and the microphone 74 may be wired or wireless, similar to the camera 71 and the like.

[0051] The storage unit 63 includes a non-volatile memory element such as a flash memory or an EEPROM (Electrically Erasable Programmable Read Only Memory). The storage unit 63 stores various programs executed by the control unit 61 and various databases (DB) required for the processing of the control unit 61. In this embodiment, the storage unit 63 stores the program 60P executed by the control unit 61. The storage unit 63 also stores an image DB 631, a training image DB 632, an audio DB 633, and a training audio DB 634, as well as a first learning model M1 and a second learning model M2.

[0052] The communication unit 64 communicates with other devices. If it is determined that the biomass fuel contains unsuitable fuel, the damper 14 of the fuel receiving unit 10, the slide gate 32 of the push-in transport unit 30, and the slide gate 41 of the fuel input unit 40 are operated by a command sent through the communication unit 64, and the unsuitable fuel is discharged.

[0053] The display unit 65 is configured with a liquid crystal panel, an organic EL (Electro Luminescence) display, or the like. The display unit 65 displays various images, characters, and the like based on the processing of the control unit 61. The operation unit 66 accepts user operations and notifies the control unit 61 of the accepted operations. For example, the operation unit 66 accepts user operations through an input device such as a mechanical button or a touch panel provided on the surface of the display unit 65. Furthermore, for example, the operation unit 66 may be an input device such as a mouse and a keyboard.

[0054] Next, the discharge mechanism for unsuitable fuel will be described. Figure 3 is an explanatory diagram showing an example of the discharge mechanism. The discharge mechanism shown in Figure 3 shows a minimum configuration. The discharge mechanism shown in Figure 3 can be applied appropriately depending on the construction of the biomass fuel transport line.

[0055] FIG. 3A shows an example of a discharge mechanism when biomass fuel is transferred by directly transferring between two conveyors. The left diagram of FIG. 3A shows a normal state, and the right diagram shows a state in which an unsuitable fuel is detected. As shown in the left diagram of FIG. 3A, in a normal state, the plate-shaped part of the damper DM is in a vertical state, and the biomass fuel being transported by the conveyor CV1 falls toward the conveyor CV2 at the right end of the conveyor CV1. The dropped biomass fuel is received by the conveyor CV2 and transported. In FIG. 3A, the biomass fuel being transported by the conveyor CV1 is monitored by the camera CM. The image taken by the camera CM is sent to a determination device such as the control device 60, which determines whether or not the biomass fuel contains unsuitable fuel. If the determination device determines that unsuitable fuel is present, the plate-shaped part of the damper DM transitions to a state of a predetermined angle, as shown in the right diagram of FIG. 3A. Materials unsuitable for combustion or biomass fuel mixed with materials unsuitable for combustion that fall from the right end of conveyor CV1 do not fall onto conveyor CV2 but are discharged.

[0056] FIG. 3B shows an example of a discharge mechanism when biomass fuel is transferred via a chute and transferred between two conveyors. The left diagram of FIG. 3B shows a normal state, and the right diagram shows a state in which an unsuitable fuel is detected. As shown in the left diagram of FIG. 3B, in a normal state, the chute SH is in a vertical state. The biomass fuel being transported by the conveyor CV1 falls into the chute SH at the right end of the conveyor CV1, is transported via the chute SH, and is received and transported by the conveyor CV2. In FIG. 3B, the biomass fuel being transported by the conveyor CV1 is monitored by a camera CM. In addition, the sound generated by the chute SH is collected by a microphone MC. The image taken by the camera CM and the sound collected by the microphone MC are sent to a determination device such as a control device 60, which determines whether or not the biomass fuel contains unsuitable fuel. If the determination device determines that an unsuitable fuel is present, the chute SH transitions to a bent plate shape as shown in the right diagram of FIG. 3B. Materials unsuitable for combustion that fall into the chute SH at the right end of the conveyor CV1, or biomass fuel mixed with materials unsuitable for combustion, are transported through the chute SH and discharged without falling onto the conveyor CV2.

[0057] In both the discharge mechanisms of Figures 3A and 3B, the camera CM monitors the vicinity of the right end of the conveyor CV1 because it can reliably detect when unsuitable for combustion reaches the right end of the conveyor CV1. This makes it possible to minimize the amount of biomass fuel discharged together with the unsuitable for combustion. In addition, if the structure of the conveying line makes it impossible to install the camera CM in a position that can monitor the right end of the conveyor CV1, the time from the position where the unsuitable for combustion is detected until it reaches the right end of the conveyor CV1 is calculated, and the damper DM or chute SH is controlled based on the calculation result.

[0058] The discharge mechanism shown in Fig. 3A can be employed in the above-mentioned fuel receiving section 10. The discharge mechanism shown in Fig. 3B can be employed in the above-mentioned supply and transfer section 20.

[0059] Next, the database stored by the control device 60 will be described. FIG. 4 is an explanatory diagram showing an example of an image DB. The image DB 631 stores images captured by a camera in association with the determination result of whether or not the image contains an unsuitable substance for fuel. The image DB 631 includes an ID column, an image column, a shooting date and time column, a camera position column, a determination column, a position column, a type column, and a certainty column. The ID column stores an ID that can identify a record. The image column stores the captured image as binary data. The image data may be stored in the storage unit 63 as a file, and the image column may store a file name. The shooting date and time column stores the date and time when the image was captured. The shooting time may store a frame number as well as a time code. The camera position column stores the position of the camera that captured the image. The determination column stores the determination result of whether or not the image contains an unsuitable substance for fuel. For example, if it is determined that an unsuitable substance for fuel is included, the determination column stores "yes," and if it is determined that an unsuitable substance for fuel is not included, the determination column stores "no." The position column stores image coordinate values ​​indicating the position of an unsuitable fuel object when it is determined that the object is included. For example, the position column stores the coordinates of the upper left corner and the coordinates of the lower right corner of a bounding box that encloses the area of ​​the unsuitable fuel object. The position column may store the center coordinates, width, and height of the object. The type column stores the type of the unsuitable fuel object when it is determined that the object is included. For example, the type may be a long object, a metal piece, a stone, etc. The certainty column stores the certainty output by the learning model. The certainty is a value that exceeds 0 and is below 1, and a larger value indicates a higher certainty.

[0060] FIG. 5 is an explanatory diagram showing an example of the training image DB. The training image DB 632 stores training images for learning (generating) or relearning the first learning model M1 described later, and correct answer data (labels) of the judgment, etc., in association with each other. The training image DB 632 includes an ID column, an image column, a shooting date column, a camera position column, a judgment column, a position column, and a type column. The contents stored in the ID column to the camera position column are the same as the contents stored in the columns of the same name in the image DB 631, so their explanations are omitted. The judgment column stores correct answer data in the judgment of whether or not an unsuitable substance for fuel is included in the image. If an unsuitable substance for fuel is included in the image, the judgment column stores "Yes". If an unsuitable substance for fuel is not included in the image, the judgment column stores "No". If an unsuitable substance for fuel is included in the image, the position column stores image coordinate values ​​indicating the position of the unsuitable substance for fuel when the unsuitable substance for fuel is included in the image. The values ​​stored in the position column are the same as the contents stored in the position column of the image DB 631, so their explanations are omitted. The type column stores the type of the unsuitable fuel object when the unsuitable fuel object is included. Specific examples of the type are the same as those in the image DB 631.

[0061] FIG. 6 is an explanatory diagram showing an example of the voice DB. The voice DB 633 stores voice collected by a microphone, features of the voice, and a judgment result of whether or not the voice includes a voice generated due to an unsuitable fuel, in association with each other. The voice DB 633 includes an ID string, a voice string, a collection date and time string, a microphone position string, a feature string, a judgment string, a type string, and a certainty string. The ID string stores an ID that can identify a record. The voice string stores the collected voice as binary data. The voice data may be stored in the storage unit 63 as a file, and a file name may be stored in the voice string. The collection date and time string stores the date and time when the voice was collected. The microphone position string stores the position of the microphone. The feature string stores a feature created from the collected voice data. The feature may be text data or binary data. The feature may be stored in the storage unit 63 as a file, and a file name may be stored in the feature string. The feature will be described later. The judgment string stores a judgment result of whether or not the voice includes a voice generated due to an unsuitable fuel. For example, if it is determined that the audio generated due to an unsuitable fuel is included, the judgment column stores "yes." If it is determined that the audio generated due to an unsuitable fuel is not included, the judgment column stores "no." If it is determined that the audio generated due to an unsuitable fuel is included, the type column stores the type of the unsuitable fuel. Examples of types are the same as those in image DB 631. The certainty column stores the certainty output by the learning model. The certainty is a value exceeding 0 and below 1, with a larger value indicating a higher certainty.

[0062] The features of speech are explained below. The features of speech are extracted from speech collected by a microphone using techniques such as frequency analysis and spectrum analysis. The features include spectrograms, Mel spectra, and Mel frequency cepstrum coefficients. These features may be expressed as multi-dimensional coordinate sequences or as images.

[0063] FIG. 7 is an explanatory diagram showing an example of the training voice DB. The training voice DB 634 stores voice features for learning (generating) or relearning the second learning model M2 described later, and correct answer data (labels) of the judgment, etc., in association with each other. The training voice DB 634 includes an ID string, a voice string, a collection date and time string, a microphone position string, a judgment string, and a type string. The contents stored in the ID string to the microphone position string are the same as the contents stored in the strings of the same name in the voice DB 633, so their explanations are omitted. The judgment string stores correct answer data in the judgment of whether or not the collected voice contains a voice generated due to the unsuitable fuel. If the collected voice contains a voice generated due to the unsuitable fuel, the judgment string stores "Yes". If the collected voice does not contain a voice generated due to the unsuitable fuel, the judgment string stores "No". If the collected voice contains a voice generated due to the unsuitable fuel, the type string stores the type of the unsuitable fuel. Specific examples of the types are the same as those in the voice DB 633.

[0064] Next, a learning model used for detecting unsuitable for combustion will be described. FIG. 8 is an explanatory diagram showing an example of the first learning model. The first learning model M1 is a learning model that judges whether or not an unsuitable for fuel is included in an image. FIG. 8A shows the first learning model M1 when judging whether or not an unsuitable for fuel is included. The first learning model M1 is a learning model that is trained to recognize an unsuitable for fuel in an image when image data (sensor data) and a camera position are input. The first learning model M1 is configured using R-CNN (Regional CNN), SSD (Single Shot MultiBox Detector), DETR (DEtection with TRansformers), YOLO (You Only Look Once), and random forest. In the following description, it is assumed that the first learning model M1 is configured using a neural network such as YOLO. When images captured by cameras 71, 72, and 73 and the positions of the respective cameras are input to the first learning model M1, the first learning model M1 outputs the presence or absence of an unsuitable for fuel. Furthermore, when the first learning model M1 determines that an unsuitable fuel object is present, it outputs the coordinates indicating the position of the unsuitable fuel object and the confidence level. The coordinates are, for example, the coordinates (Xmax, Ymax) and (Xmin, Ymin) of the bounding box surrounding the unsuitable fuel object.

[0065] FIG. 8B shows the first learning model M1 at the time of generation or re-learning. Images and camera positions included in the training data are input to the first learning model M1. The first learning model M1 outputs the presence or absence of unsuitable fuel objects, the coordinates (Xmax, Ymax) and (Xmin, Ymin) of the bounding box surrounding the unsuitable fuel object, the type of unsuitable fuel object, and the confidence level for the judgment result. The output value of the first learning model M1 is compared with the values ​​of the judgment column, position column, and type column included in the training data, and the first learning model M1 is trained so that the output value of the first learning model M1 approaches the values ​​of the corresponding training data.

[0066] FIG. 9 is a flowchart showing an example of the procedure of the first generation process. The first generation process is a process performed when generating or relearning the first learning model M1. The control unit 61 of the control device 60 acquires training data from the training image DB 632 (step S1). The control unit 61 selects one record from the training data (step S2). The control unit 61 performs learning using one record of the selected training data (step S3). The control unit 61 inputs the image and camera position included in the training data to the first learning model M1. The first learning model M1 outputs the presence or absence of an unsuitable fuel, the coordinates (Xmax, Ymax) and (Xmin, Ymin) of the bounding box surrounding the unsuitable fuel, the type of the unsuitable fuel, and the confidence level for the judgment result. The control unit 61 compares these values ​​with the values ​​of the judgment column, the position column, and the type column included in the training data, and optimizes parameters such as weights between neurons constituting the first learning model M1 so that the output value of the first learning model M1 approaches the value of the corresponding training data. The control unit 61 determines whether there are any unprocessed records that have not been used in learning (step S4). If the control unit 61 determines that there are any unprocessed records (YES in step S4), the process returns to step S2 and continues. If the control unit 61 determines that there is no unprocessed training data (NO in step S4), the control unit 61 stores the first learning model M1 obtained by learning in the storage unit 63 (step S5) and ends the process. Note that the first generation process (learning) may be executed by a computer other than the control device 60, and the result may be stored in the storage unit 63 of the control device 60.

[0067] FIG. 10 is an explanatory diagram showing an example of the second learning model. The second learning model M2 is a learning model that judges whether or not the collected voice contains a voice generated due to an unsuitable fuel. FIG. 10A shows the second learning model M2 when judging whether or not the collected voice contains a voice generated due to an unsuitable fuel. The second learning model M2 is a learning model that is trained to recognize a voice generated due to an unsuitable fuel in the collected voice when a voice feature (sensor data) and a microphone position are input. The second learning model M2 is configured using an autoencoder (AE), a Gaussian mixture model (GMM), or the like. When the voice feature is expressed by an image, the second learning model may be configured using an R-CNN (Regional CNN), or the like. The voice collected by the microphone 74 is subjected to an analysis process using a method such as frequency analysis or spectrum analysis to extract a feature. When the extracted feature and the microphone position are input to the second learning model M2, the second learning model M2 outputs the presence or absence of an unsuitable fuel. Furthermore, when the second learning model M2 determines that an unsuitable fuel is present, it outputs the type of the unsuitable fuel and the confidence level.

[0068] FIG. 10B shows the second learning model M2 at the time of generation or re-learning. Analysis processing is performed on the voice data included in the training data to extract features. The extracted features and the microphone positions included in the training data are input to the second learning model M2. The second learning model M2 outputs the presence or absence of unsuitable fuel, the type of unsuitable fuel, and the confidence level for the judgment result. The output value of the second learning model M2 is compared with the values ​​of the judgment sequence and the type sequence included in the training data, and the second learning model M2 is trained so that the output value of the second learning model M2 approaches the value of the corresponding training data.

[0069] FIG. 11 is a flowchart showing an example of the procedure of the second generation process. The second generation process is a process performed when generating or relearning the second learning model M2. The control unit 61 of the control device 60 acquires training data from the training voice DB 634 (step S11). The control unit 61 selects one record from the training data (step S12). The control unit 61 acquires voice data included in one record of the selected training data, performs analysis processing, and extracts features (step S13). The control unit 61 performs learning using the extracted features and the camera positions included in the selected training data (step S14). The control unit 61 inputs the extracted features and the camera positions included in the selected training data to the second learning model M2. The second learning model M2 outputs the presence or absence of an unsuitable fuel substance, the type of the unsuitable fuel substance, and the confidence level for the judgment result. The control unit 61 compares these values ​​with the values ​​of the judgment sequence and the type sequence included in the training data, and optimizes parameters such as weights between neurons and nodes constituting the second learning model M2 so that the output value of the second learning model M2 approaches the value of the corresponding training data. The control unit 61 determines whether there is an unprocessed record that has not been used in learning (step S15). If the control unit 61 determines that there is an unprocessed record (YES in step S15), the process returns to step S12 and continues. If the control unit 61 determines that there is no unprocessed training data (NO in step S15), the control unit 61 stores the second learning model M2 obtained by learning in the storage unit 63 (step S16), and ends the process. The second generation process (learning) may be executed by a computer other than the control device 60, and the result may be stored in the storage unit 63 of the control device 60.

[0070] FIG. 12 is a flow chart showing an example of the procedure of the judgment process. The judgment process is a process for judging whether or not the biomass fuel contains any unsuitable fuel. The control unit 61 of the control device 60 acquires images from the cameras 71, 72, and 73 (step S31). The acquired images are videos of a predetermined length or a predetermined number of still images. Since a fixed amount of biomass fuel is discharged at a fixed interval by the fixed quantity feeder 13, the shutters of the cameras 71, 72, and 73 may be released accordingly to acquire still images. For example, a photoelectric sensor may be provided to detect whether or not the biomass fuel is present on the conveyor, and the shutter may be released based on the detection result of the photoelectric sensor. The control unit 61 judges whether or not the unsuitable fuel is present for each acquired image (step S32). The control unit 61 inputs the acquired image together with information on the camera position to the first learning model M1 and acquires the judgment result. In this embodiment, since there are three cameras, at least three images are acquired and judgment is performed at least three times. The control unit 61 acquires voice from the microphone 74 (step S33). The acquired voice is a digital voice of a predetermined length. As in the case of image acquisition, the timing of voice acquisition may be adjusted according to the flow of biomass fuel. In this embodiment, since it is possible to detect that biomass fuel is being put into the chute 23 by the image captured by the camera 72, voice may be acquired from the microphone 74 at the timing of the input. The control unit 61 judges whether or not an unsuitable fuel is mixed in for each acquired voice (step S34). The control unit 61 inputs the acquired voice together with information on the position of the microphone 74 to the second learning model M2 and acquires the judgment result. Since there is one microphone in this embodiment, one voice is acquired and one judgment is performed. The control unit 61 judges whether or not an unsuitable fuel is mixed in based on the judgment results of the first learning model M1 and the second learning model M2 (step S35). Here, the cameras 71, 72, 73, and the microphone 74 are each monitoring a different biomass fuel at a different location, so that judgment is made individually. If the control unit 61 determines that no unsuitable fuel substance has been mixed in at any of the judgment points (NO in step S35), the process proceeds to step S37. If the control unit 61 determines that any unsuitable fuel substance has been mixed in at any of the judgment points (YES in step S35), the control unit 61 transmits a discharge command (step S36).The control unit 61 transmits an operation command to a discharge mechanism, a damper, a chute, a slide gate, etc. according to the determined location, and discharges the biomass fuel determined to contain unsuitable fuel to the outside. The control unit 61 stores the determination result (step S37). The control unit 61 associates the image used in the determination with the determination result and stores them in the image DB 631. The control unit 61 also associates the voice data used in the determination with the determination result and stores them in the voice DB 633. The control unit 61 determines whether or not to end the determination process (step S38). If the control unit 61 determines not to end the determination process (NO in step S38), it returns the process to step S31. If the control unit 61 determines to end the determination process (YES in step S38), it ends the process. The determination process continues to be executed without ending while the fuel supply system 100 is operating. As described above, when the acquisition of images and sounds is performed intermittently by adjusting the timing, the determination process may be started and ended every time an image or sound is acquired.

[0071] The first learning model M1 is assumed to determine the presence or absence of unsuitable fuel materials for images captured by all of the cameras 71, 72, and 73, but this is not limited to the above. A first learning model M1 may be provided for each camera. In this case, the camera position is not required as an input to the first learning model M1. The same is true for the second learning model M2. When multiple microphones are installed and a second learning model M2 is provided for each microphone, the microphone position is not required as an input to the second learning model M2.

[0072] When the inclusion of an unsuitable fuel is detected, the unsuitable fuel is discharged by a discharge mechanism located after the detection point. If the control of the discharge mechanism is not in time and the unsuitable fuel cannot be discharged, the unsuitable fuel is discharged by a discharge mechanism located further back. The biomass fuel containing the unsuitable fuel discharged by the discharge mechanism may be photographed by a camera, and the photographed image may be used to determine whether the unsuitable fuel has been discharged. If the final discharge mechanism fails to discharge the unsuitable fuel, the operation of the fuel supply system 100 may be stopped, and the unsuitable fuel may be removed manually. When the inclusion of a long object as an unsuitable fuel is detected, if the long object may get caught in a conveying mechanism such as a conveyor or a chute before it reaches the discharge mechanism and may affect the operation of the conveying mechanism, the operation of the fuel supply system 100 may be stopped, and the long object may be removed manually.

[0073] If the amount of biomass fuel introduced into the boiler combustion chamber 80 is insufficient as a result of discharging biomass fuel contaminated with unsuitable fuel, the control device 60 may control the quantitative feeder 13, etc., to temporarily increase the amount of biomass fuel supplied to the pushing conveying section 30. Whether or not the amount of biomass fuel is insufficient is determined based on the amount of biomass fuel returned by the circulating conveying section 50, the combustion state obtained by a camera monitoring the boiler combustion chamber 80, and the temperature and pressure inside the boiler combustion chamber obtained by various sensors.

[0074] In this embodiment, it is possible to detect unsuitable fuels from a mixture containing biomass fuel. In addition, when unsuitable fuels are mixed into biomass fuel, the mixture containing the unsuitable fuels is discharged by the transport mechanism, so that the unsuitable fuels can be discharged without stopping the operation of the fuel supply system 100. As a result, the fuel supply system 100 can be operated without constant monitoring by a person such as an observer.

[0075] (aggregation function) Next, the additional functions will be described. The function of counting the unsuitable fuel substances will be described. FIG. 13 is a flow chart showing an example of the procedure of the counting process. The control unit 61 of the control device 60 acquires the range to be counted (step S51). For example, an operator inputs the period (start date and end date) for which counting is to be performed to the operation unit 66. The operation unit 66 sends the input period to the control unit 61, and the control unit 61 acquires the period as the counting range. The control unit 61 acquires data corresponding to the counting period from the image DB 631 and the voice DB 633 (step S52). The control unit 61 performs overlap check (step S53). The overlap check is performed because there may be cases where the same unsuitable fuel substances are detected by multiple cameras, or where the same unsuitable fuel substances are detected by a camera and a microphone. The overlap check is performed, for example, by date and time and type of unsuitable fuel substances. If there is overlap, the control unit 61 keeps only one piece of overlapping data and excludes the rest from the counting target. The control unit 61 counts the number of unsuitable fuel items by type (step S54). The control unit 61 outputs the counting result (step S55) and ends the process. The output of the counting result may be displayed on the display unit 65 or may be stored in the storage unit 63.

[0076] 14 is an explanatory diagram showing an example of a tally result display screen. The tally result display screen d01 includes a period display d011, a case count table d012, and a close button d013. The period display d011 indicates the tally period. The case count table d012 indicates the number of cases of each type of unsuitable fuel. When the close button d013 is selected, the tally result display screen is closed. The tally result display screen may be displayed on a terminal capable of communicating with the control device 60.

[0077] The aggregation function makes it possible to check the status of contamination by unsuitable fuel materials. In addition, by adjusting the scope of aggregation, it is possible to estimate the status of contamination by unsuitable fuel materials for each lot of biomass fuel. This makes it possible to feed back the results to the biomass fuel delivery destination and request improvements.

[0078] (Image and audio confirmation function) The image / audio confirmation function is a function for confirming an image or audio that the control device 60 has determined to include an unsuitable fuel substance. FIG. 15 is a flow chart showing an example of a procedure for confirmation screen display processing. The confirmation screen display processing is processing for displaying a screen including an image or the like including an unsuitable fuel substance. The control unit 61 of the control device 60 acquires data on the detected unsuitable fuel substance from the image DB 631 and the audio DB 633 (step S71). It is desirable to narrow down the number of data to be acquired to some extent, such as the most recent 50 cases or data within the most recent 12 hours. The control unit 61 performs correspondence between images or audio that detect the same unsuitable fuel substance in the acquired data (step S72). For example, if the type is the same and the detection time is close, it is determined that they correspond. The control unit 61 creates a list of the detected unsuitable fuel substances based on the acquired data and the result of the correspondence, and outputs it to the display unit 65 (step S73). The operator refers to the list and performs operation input such as selecting the unsuitable substance that the operator wants to confirm details about. The control unit 61 receives an operation input (step S74). The control unit 61 judges whether the received input is a record designation (selection of an inappropriate object) or not (step S75). If the control unit 61 judges that the input is a record designation (YES in step S75), it creates and outputs a confirmation screen (step S76). The operator checks the confirmation screen and performs an operation input. The control unit 61 receives the operation input (step S77). The control unit 61 judges whether the received input is a selection of a voice icon or not (step S78). If the control unit 61 judges that the received input is a selection of a voice icon (YES in step S78), it plays a voice (step S79) and returns the process to step S77. If the control unit 61 judges that the received input is not a selection of a voice icon (NO in step S78), it judges whether or not to return to the list (step S80). If the control unit 61 determines that the received input is "return to list" (YES in step S80), the process returns to step S73. If the control unit 61 determines that the received input is not "return to list" (NO in step S80), the process returns to step S77. If the control unit 61 determines that the input is not a record designation (NO in step S75), the control unit 61 determines whether to close the list (step S81).When the control unit 61 determines not to close the list (NO in step S81), the process returns to step S74. When the control unit 61 determines to close the list (YES in step S81), the control unit 61 closes the list screen (step S82) and ends the process.

[0079] FIG. 16 is an explanatory diagram showing an example of a list screen. The list screen d02 includes a list table d021 and a close button d022. The list table d021 includes a date column, a time column, a type column, a camera 1 column, a camera 2 column, a camera 3 column, and a microphone 1 column. The date column displays the date when the unsuitable fuel object was detected. The time column displays the time when the unsuitable fuel object was detected. If the same unsuitable fuel object was detected multiple times by multiple cameras, for example, the first time is displayed. The type column displays the type of unsuitable fuel object detected. In the camera 1 column, an icon is displayed when the object is detected using an image captured by the camera 71. If not, a predetermined symbol, for example, "-", is displayed. The same applies to the camera 2 column and the camera 3 column. In the microphone 1 column, an icon is displayed when the object is detected using a voice collected by the microphone 74. The number displayed under the icon indicates the confidence level output by the learning model. The confidence level is a value between 0 and 1, and the higher the value, the higher the confidence level. When an operation to select one record is performed in the list, a confirmation screen is displayed. When the close button d022 is selected, the list screen closes.

[0080] FIG. 17 is an explanatory diagram showing an example of a confirmation screen. The confirmation screen d03 includes a date and time display d031, a type display d032, a location display d033, an image area d034, an audio icon d035, and a back button d036. The date and time display d031 indicates the date and time when the unsuitable fuel object was detected. The type display d032 indicates the type of the unsuitable fuel object. The location display d033 indicates the location where the unsuitable fuel object was detected. The image area d034 displays an image including the unsuitable fuel object (unsuitable object image). In the image displayed in the image area d034, the detected unsuitable fuel object is surrounded by a bounding box (figure). The type of the unsuitable fuel object may be displayed in a display form such as the line color or line shape of the bounding box. In the example of FIG. 17, the type is indicated as stone by a dotted line. When detection is performed by multiple cameras, multiple image areas d034 are displayed. When the audio icon d035 is selected, audio is played when the unsuitable fuel object is detected. If no fuel unsuitable materials are detected by voice, the voice icon d035 is not displayed. When the back button d036 is selected, the screen returns to the list screen.

[0081] The image and audio confirmation function makes it possible to confirm the unsuitable fuel that has actually been detected. In particular, when operations have to be stopped temporarily to manually remove the unsuitable fuel, being able to refer to the image of the unsuitable fuel is expected to improve the efficiency of the work.

[0082] (Implementation support) Next, support for the introduction of a removal device for unsuitable fuel materials will be explained. The removal device is a device that detects and discharges unsuitable fuel materials. In the above example, the removal device corresponds to a set including the control device 60, the cameras 71, 72, 73, and the microphone 74, as well as a discharge mechanism such as the damper 14. The target of the introduction support is biomass fuel processing facilities at biomass fuel power plants that do not have a mechanism for detecting and discharging unsuitable fuel materials.

[0083] The flow of the introduction support will be described. FIG. 18 is a flow chart showing the flow of the introduction support work. First, in the biomass fuel processing facility to be introduced, the supply route from the receipt of biomass fuel to the boiler combustion chamber is confirmed (step S91). Based on the confirmation result, the temporary installation positions of the cameras and microphones are determined (step S92). The cameras and microphones are temporarily installed at the determined installation positions (step S93). The determination processing program for the unsuitable fuel and the first learning model and the second learning model are introduced to the control device that controls the supply of biomass fuel (step S94). Note that it is preferable to use trained models that have already been used in actual operations of other facilities, etc., as the first learning model and the second learning model. Even when a trained model is used, the first learning model may be retrained using an image obtained from a temporarily installed camera. Similarly, the second learning model may be retrained using a voice obtained from a microphone. The inclusion of unsuitable fuel is monitored by the cameras and microphones (step S95). After a predetermined time or after a predetermined time has elapsed, the types and numbers of unsuitable fuel detected are tallied (step S96). Based on the results of the compilation and the installation positions (mounting positions) of the cameras and microphones, the installation location of the discharge mechanism, etc. is considered (step S97). The cameras, microphones, and discharge mechanism are actually installed (step S98). The settings of the judgment processing program are adjusted according to the installation status of the cameras, microphones, and discharge mechanism (step S99). Finally, an operational test is performed (step S100) to complete the work. If the results of the operational test are unsatisfactory, the installation positions of the cameras, microphones, and discharge mechanism are reconsidered, etc.

[0084] By carrying out the introduction support work in accordance with the flow shown in Figure 18, it becomes possible to smoothly introduce the fuel unsuitable material removal equipment to biomass fuel processing facilities, etc.

[0085] The technical features (constituent elements) described in each embodiment can be combined with each other, and by combining them, new technical features can be formed. The embodiments disclosed herein are illustrative in all respects and should not be considered as limiting. The scope of the present invention is indicated by the claims, not by the above meaning, and is intended to include all modifications within the meaning and scope of the claims. In addition, the claims use a format in which a claim cites two or more other claims (multiple claim format), but are not limited to this. A format in which a multiple claim cites at least one multiple claim (multiple multiple claim) may also be used. [Explanation of symbols]

[0086] 100: Fuel supply system 10: Fuel receiving section 20: Supply and conveying section 30: Push-in transport section 40: Fuel input section 50: Circulation conveying section 60: Control device 60P: Program 61: Control section 62: Input / output section 63: Storage section 631: Image DB 632: Training image DB 633: Audio Database 634: Training voice database M1: First learning model M2: Second learning model 64: Communications Department 65:Display section 66:Operation unit 70: Monitoring Department 80: Boiler combustion chamber

Claims

1. a sensor for detecting a mixture including a biomass fuel; A control unit. The control unit detects fuel unsuitable substances in the mixture based on the sensor data acquired by the sensor. Biomass fuel unsuitable material detection system.

2. the sensor is a camera that photographs the mixture, The control unit is acquiring mixture image data including the mixture; When mixture image data is input, the acquired mixture image data is input to a learning model that has been trained to recognize unsuitable fuel materials, including long objects and metals, and the detection results of unsuitable fuel materials are output. The biomass fuel unsuitable substance detection system according to claim 1 .

3. The camera is installed at the monitoring port of the quantitative feeder. The biomass fuel unsuitable substance detection system according to claim 2.

4. Count the number of unsuitable fuel items detected by type. The biomass fuel unsuitable substance detection system according to any one of claims 1 to 3.

5. An image of the unsuitable material including the detected unsuitable material for fuel is output. The biomass fuel unsuitable substance detection system according to any one of claims 1 to 3.

6. The image of the unsuitable material has a figure showing an area of ​​the unsuitable material for fuel drawn thereon. The biomass fuel unsuitable substance detection system according to claim 5.

7. The graphic representation is different depending on the attributes including the type of fuel unsuitable substance. The biomass fuel unsuitable substance detection system according to claim 6.

8. the sensor is a microphone; The control unit is Acquire mixture sound data generated when the mixture moves; When the mixture voice data is input, the acquired mixture voice data is input to a second learning model that is trained to recognize unsuitable fuel materials including long objects and metals, and the presence or absence of unsuitable fuel materials is output. The biomass fuel unsuitable substance detection system according to any one of claims 1 to 3.

9. The microphone is installed in a chute that constitutes a path through which the mixture falls. The biomass fuel unsuitable substance detection system according to claim 8.

10. When the unsuitable fuel is detected in the chute, the unsuitable fuel is re-identified on the conveyor on which it is being transported after being discharged from the chute; The unsuitable fuel is discharged by switching the damper of the conveyor, or the unsuitable fuel is discharged together with the surplus biomass fuel by closing the supply port of the biomass fuel. The biomass fuel unsuitable substance detection system according to claim 9.

11. Outputs audio data of fuel inappropriate items along with image data The biomass fuel unsuitable substance detection system according to claim 8.

12. A sensor detects a mixture containing biomass fuel; Based on the sensor data acquired by the sensor, fuel unsuitable substances in the mixture are detected. A program for detecting unsuitable biomass fuels that uses a computer to perform the processing.

13. A sensor detects a mixture containing biomass fuel; Based on the sensor data acquired by the sensor, fuel unsuitable substances in the mixture are detected. A method for detecting unsuitable materials for biomass fuel, in which processing is performed by a control device.

14. A camera is installed in a biomass fuel processing facility at a biomass fuel power plant that does not have a device for removing unsuitable fuel materials, and image data of the mixture containing biomass fuel is obtained. The acquired mixture image data is input to a learning model that has been trained to recognize unsuitable fuel materials including long objects and metals when the mixture image data is input, and the unsuitable fuel materials are output by inputting the acquired mixture image data; outputting the type and number of the unsuitable fuel objects; The fuel unsuitable substance removal device is installed based on the mounting position of the camera. Installation method.

Images