JP2023127844A - Information processing device, information processing method, and information processing program - Google Patents

Abstract

To provide a technology to classify, predict, and estimate a combustion state in an incinerator with a higher degree of precision.SOLUTION: An information processing device includes: a first image analysis unit for executing evaluation in a first evaluation axis with new first imaging data on the inside of an incinerator as an input using a first learned model that has mechanically learned first teacher data generated by granting a classification label in the first evaluation axis, which serves as an element for determining a combustion state, to the first imaging data imaging the inside of the incinerator; a second image analysis unit for executing evaluation in a second evaluation axis with new second imaging data on the inside of the incinerator as an input using a second learned model that has mechanically learned second teacher data generated by granting a classification label in the second evaluation axis, which serves as an element for determining a combustion state, to the second imaging data that captures the inside of the incinerator; and a combustion state determination unit for determining a present combustion state by mapping an evaluation result in each of the first evaluation axis and the second evaluation axis on a combustion state determination map.SELECTED DRAWING: Figure 2

Description

The present invention relates to an information processing device, an information processing method, and an information processing program for determining a combustion state in an incinerator.

Waste incineration facilities not only efficiently recover heat from combustion exhaust gas, but also suppress the emission of harmful substances (CO, dioxins, NOx, etc.).In order to stabilize the combustion of waste, automatic Combustion control (ACC) has been introduced.

However, if the ACC cannot maintain the combustion state in the incinerator due to significant fluctuations in waste quality, a skilled operator may intervene manually. At this time, the skilled operator makes a decision on manual intervention based on process values obtained from various sensors and combustion images taken of the combustion state.

The information from the above combustion images is a very important indicator, and various companies have applied for patents for technology that classifies combustion conditions and predicts process values from combustion images, mainly using image recognition technology using deep learning. . For example, Patent Document 1 proposes the use of a classification model that uses combustion images as input to classify combustion states into, for example, eight ways. Further, Patent Document 2 proposes creating an estimation model that uses combustion images as input to predict estimated values (CO concentration, NOx concentration, unburned content in ash, etc.) according to the combustion state.

JP2021-8991 Patent No. 6824859

In Patent Document 2, a combustion image is linked to measured values obtained as a result of combustion (CO concentration, NOx concentration, unburned content in ash, etc.), and a model is created to predict the measured values from the combustion image. There is. However, since the above measured values are obtained as a result of combustion conditions and conditions that change from moment to moment, it is better to accurately understand the combustion conditions than to predict the above measured values. It is considered to be an important element in status understanding and subsequent operation management (combustion control).

In Patent Document 1, combustion image data is used as an input parameter for learning, and data quantifying the strength of the combustion state, data quantifying the position of the burnout point of combustion, and the degree of generation of unburned matter are quantified. A model is generated by machine learning of training data using data as output parameters for learning. However, Patent Document 1 describes that the trained model classifies the combustion state into multiple ways (e.g. 8 ways), but the details of the relationship between the digitized data that are output parameters and the classification are This is not clearly stated, and there is a possibility that the combustion state cannot be classified properly.

The present invention has been made in consideration of the above points. An object of the present invention is to provide a technology for classifying, predicting, and estimating the combustion state in an incinerator with higher accuracy.

The information processing device according to the first aspect of the present invention includes:
First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a first image analysis unit that performs evaluation on the first evaluation axis by inputting new first imaging data in the incinerator using a first learned model that has been machine-learned;
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. a second image analysis unit that performs evaluation on the second evaluation axis by inputting new second imaging data in the incinerator using a second learned model that has been machine-learned;
By mapping the evaluation results on each of the first evaluation axis and the second evaluation axis onto a predetermined combustion state determination map whose coordinate axes are the first evaluation axis and the second evaluation axis, the current combustion state can be determined. a combustion state determination unit that determines the state;
Equipped with.

According to such an aspect, the first imaging data inside the incinerator is used to perform evaluation on the first evaluation axis, which is an element for determining the combustion state, and the second imaging data inside the incinerator is used to perform evaluation on the first evaluation axis, which is an element for determining the combustion state. , the evaluation is performed on the second evaluation axis, which is an element for determining the combustion state, and the combustion state is determined based on the evaluation results on each of the first evaluation axis and the second evaluation axis. Compared to the case where the combustion state is determined based on the evaluation results, the combustion state can be classified, predicted, and estimated with higher accuracy. In addition, since the combustion state is determined by mapping the evaluation results on each of the first evaluation axis and the second evaluation axis onto the combustion state determination map, the evaluation results on each of the first evaluation axis and the second evaluation axis are It is possible to intuitively grasp the relationship between the combustion state and the combustion state that is the determination result, and it is also possible to quickly determine the combustion state.

An information processing device according to a second aspect of the present invention is an information processing device according to the first aspect, comprising:
The first imaging device is an infrared imaging device, and the second imaging device is a visible light imaging device.

According to this aspect, the visible light imaging device can acquire an image that allows the situation of the flame to be grasped as the second imaging data (visible light imaging data) inside the incinerator, whereas the infrared imaging device By appropriately selecting the imaging wavelength, the first imaging data (infrared imaging data) inside the incinerator can be used to capture images from the upstream side of the incinerator, where the influence of flames has been removed, to the image of the actually burning objects. It is possible to obtain an image containing. Therefore, infrared imaging data inside the incinerator is used to evaluate the first evaluation axis, which is an element for determining the combustion state, and visible light imaging data inside the incinerator is used to evaluate the combustion state. By performing evaluation on the second evaluation axis and using the combined evaluation results to determine the combustion state, it is possible to classify, predict, estimate, and process the combustion state from only combustion images that can only grasp the flame situation. Compared to the case of predicting and estimating values, the combustion state in the incinerator can be classified, predicted, and estimated with higher accuracy.

The information processing device according to the third aspect of the present invention is the information processing device according to the second aspect, comprising:
The first evaluation axis is at least one of the following: the amount of materials to be burned, the quality of materials to be burned, the type of materials to be burned, the temperature of materials to be burned, the amount of combustion gas generated (CO, etc.), and the temperature of the wall of the incinerator. Contains one.

The information processing device according to the fourth aspect of the present invention is the information processing device according to the second or third aspect,
The second evaluation axis includes the flame condition, the position of the combustion completion point, the shape of the combustion completion point, the amount of combustible material, the quality of combustible material, the type of combustible material, the amount of unburned material, and the amount of incinerated ash. , including at least one of the following.

An information processing device according to a fifth aspect of the present invention is an information processing device according to any one of the first to fourth aspects, comprising:
The first imaging data is moving image data within 60 seconds, and/or
The second imaging data is moving image data within 60 seconds.

According to the findings of the present inventors, the combustion state within the incinerator changes every 5 to 10 seconds. Therefore, according to this aspect, by using moving image data within 60 seconds as one segment as infrared imaging data inside the incinerator, the combustion state inside the incinerator can be classified, predicted, and estimated with higher accuracy. be able to.

An information processing device according to a sixth aspect of the present invention is an information processing device according to any one of the first to fifth aspects,
The first imaging data is moving image data of 5 seconds or more, and/or
The second imaging data is moving image data of 5 seconds or more.

According to this aspect, while the combustion state inside the incinerator changes every 5 to 10 seconds, by using video data of 5 seconds or more as one segment as infrared imaging data inside the incinerator, The combustion state inside the incinerator can be classified, predicted, and estimated with higher accuracy.

An information processing device according to a seventh aspect of the present invention is an information processing device according to any one of the first to sixth aspects,
The classification label in the first training data is at least one of a label indicating which of predetermined classification items it falls under, and a relative order among the plurality of first imaging data, and /or,
The classification label in the second teacher data is at least one of a label indicating which of predetermined classification items it falls under, and a relative order among the plurality of second imaging data.

As a result of actual verification by the inventors of the present invention, machine learning was performed using training data generated by assigning a relative order on the evaluation axis among multiple pieces of imaged data to the imaged data inside the incinerator. (Ranking learning) By performing an evaluation on the relevant evaluation axis using a trained model and determining the combustion state based on the evaluation result, the combustion state in the incinerator can be classified, predicted, and estimated with higher accuracy. was confirmed. Therefore, according to such an aspect, the combustion state in the incinerator can be classified, predicted, and estimated with higher accuracy.

An information processing device according to an eighth aspect of the present invention is an information processing device according to any one of the first to seventh aspects, comprising:
The first image analysis unit performs a combination of first imaging data in the incinerator and process data obtained from a sensor installed in the facility and/or a calculation amount obtained by calculation from the process data. New first imaging data in the incinerator and new process data are generated using a first learned model obtained by machine learning the first teaching data generated by assigning a classification label in the first evaluation axis. and/or performs evaluation on the first evaluation axis by inputting a combination with the amount of calculation obtained by calculation from the process data, and/or
The second image analysis unit performs a combination of second imaging data in the incinerator and process data obtained from a sensor installed in the facility and/or a calculation amount obtained by calculation from the process data. New second imaging data in the incinerator and new process data are generated using a second learned model obtained by machine learning the second training data generated by assigning classification labels in the second evaluation axis. And/or the combination with the calculation amount obtained by calculation from the process data is input to perform evaluation on the second evaluation axis.

An information processing device according to a ninth aspect of the present invention is an information processing device according to an eighth aspect, comprising:
The process data is data that takes into consideration a time lag between the imaging time of the first imaging data and/or the second imaging data and the response time of the sensor. The time lag may be appropriately determined based on at least one of the response speed of the sensor, the installation position of the sensor within the facility, experiments, simulations, and operator's rules of thumb.

An information processing device according to a tenth aspect of the present invention includes:
The first imaging data inside the incinerator captured by the first imaging device is given a classification label in each of two or more first evaluation axes, which are elements for determining the combustion state. a first image analysis unit that performs evaluation on each of the two or more first evaluation axes by inputting new imaging data in the incinerator using a first learned model obtained by machine learning one teacher data;
The current combustion state can be determined by mapping the evaluation results on each of the two or more first evaluation axes onto a predetermined combustion state determination map that uses the two or more first evaluation axes as coordinate axes. a combustion state determination unit that determines;
Equipped with.

According to this aspect, the first imaging data inside the incinerator is used to perform evaluation on each of the two or more first evaluation axes that are factors for determining the combustion state, and the two or more first Since the combustion state is determined based on the evaluation results for each evaluation axis, it is possible to classify, predict, and estimate the combustion state more accurately than when determining the combustion state based on the evaluation results for a single evaluation axis. can. In addition, since the combustion state is determined by mapping the evaluation results for each of two or more 12 evaluation axes onto the combustion state determination map, the evaluation results and determination results for each of the two or more first evaluation axes are determined. It is possible to intuitively grasp the relationship between the combustion state and the combustion state, and it is also possible to quickly determine the combustion state.

An information processing device according to an eleventh aspect of the present invention is an information processing device according to a tenth aspect, comprising:
The first imaging device is an infrared imaging device.

According to this aspect, by appropriately selecting the imaging wavelength, the infrared imaging device obtains the first imaging data (infrared imaging data) inside the incinerator from the upstream side of the incinerator, where the influence of the flame has been removed. Since it is possible to obtain images that include images of the burnt materials that are being burned, we use infrared imaging data inside the incinerator to perform evaluations on the first evaluation axis, which is an element for determining the combustion state. By determining the combustion state using the evaluation results, it is possible to better understand the combustion state in the incinerator compared to classifying the combustion state or predicting process values only from combustion images that can only grasp the flame situation. It is possible to classify, predict, and estimate with high accuracy.

An information processing device according to a twelfth aspect of the present invention is an information processing device according to the tenth aspect, comprising:
The first imaging device is a visible light imaging device.

An information processing device according to a thirteenth aspect of the present invention is an information processing device according to any one of the first to twelfth aspects, comprising:
The combustion state determination section corrects the combustion state determination result according to process data obtained from a sensor installed in the facility and/or the amount of calculation obtained by calculation from the process data.

An information processing device according to a fourteenth aspect of the present invention is an information processing device according to any one of the first to thirteenth aspects,
The combustion state determination section displays or issues an alert depending on the combustion state determination result.

An information processing device according to a fifteenth aspect of the present invention is an information processing device according to any one of the first to fourteenth aspects,
The apparatus further includes an instruction section that transmits an operation instruction to a crane control device and/or a combustion control device based on the determination result of the combustion state determination section.

An information processing device according to a sixteenth aspect of the present invention is an information processing device according to any one of the first to fifteenth aspects,
The algorithms used for the machine learning include maximum likelihood classification, Boltzmann machine, neural network (NN), support vector machine (SVM), Bayesian network, sparse regression, decision tree, statistical estimation using random forest, and boosting. , reinforcement learning, and deep learning.

An information processing device according to a seventeenth aspect of the present invention is an information processing device according to any one of the first to ninth aspects, comprising:
First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a first model construction unit that generates the first learned model by machine learning; and/or
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. The apparatus further includes a second model construction unit that generates the second learned model by machine learning.

The system according to the eighteenth aspect of the present invention includes:
An information processing device according to a fifteenth aspect,
The crane control device controls a crane that stirs or transports waste based on an operation instruction transmitted from the information processing device, and/or the combustion control device controls combustion of waste in an incinerator. and,
Equipped with

An information processing method according to a nineteenth aspect of the present invention is an information processing method executed by a computer, comprising:
First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a step of performing an evaluation on the first evaluation axis by inputting new first imaging data in the incinerator using a first learned model that has been machine-learned;
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. a step of performing an evaluation on the second evaluation axis by inputting new second imaging data in the incinerator using a second learned model that has been machine-learned;
By mapping the evaluation results on each of the first evaluation axis and the second evaluation axis onto a predetermined combustion state determination map whose coordinate axes are the first evaluation axis and the second evaluation axis, the current combustion state can be determined. a step of determining the state;
including.

An information processing method according to a twentieth aspect of the present invention is an information processing method executed by a computer, comprising:
The first imaging data inside the incinerator captured by the first imaging device is given a classification label in each of two or more first evaluation axes, which are elements for determining the combustion state. a step of performing evaluation on each of the two or more first evaluation axes by inputting new first imaging data in the incinerator using a first learned model obtained by machine learning one teacher data;
The current combustion state can be determined by mapping the evaluation results on each of the two or more first evaluation axes onto a predetermined combustion state determination map that uses the two or more first evaluation axes as coordinate axes. a step of determining;
including.

The information processing program according to the twenty-first aspect of the present invention causes a computer to:
First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a step of performing an evaluation on the first evaluation axis by inputting new first imaging data in the incinerator using a first learned model that has been machine-learned;
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. a step of performing an evaluation on the second evaluation axis by inputting new second imaging data in the incinerator using a second learned model that has been machine-learned;
By mapping the evaluation results on each of the first evaluation axis and the second evaluation axis onto a predetermined combustion state determination map whose coordinate axes are the first evaluation axis and the second evaluation axis, the current combustion state can be determined. a step of determining the state;
Execute.

The information processing program according to the twenty-second aspect of the present invention causes a computer to:
The first imaging data inside the incinerator captured by the first imaging device is given a classification label in each of two or more first evaluation axes, which are elements for determining the combustion state. a step of performing evaluation on each of the two or more first evaluation axes by inputting new first imaging data in the incinerator using a first learned model obtained by machine learning one teacher data;
The current combustion state can be determined by mapping the evaluation results on each of the two or more first evaluation axes onto a predetermined combustion state determination map that uses the two or more first evaluation axes as coordinate axes. a step of determining;
Execute.

According to the present invention, the combustion state in the incinerator can be classified, predicted, and estimated with higher accuracy.

FIG. 1 is a schematic diagram showing the configuration of an incineration facility according to a first embodiment. FIG. 2 is a block diagram showing the configuration of the information processing device according to the first embodiment. FIG. 3 is a flowchart illustrating an example of an information processing method by the information processing apparatus according to the first embodiment. FIG. 4 is a diagram showing an example of a combustion state determination map. FIG. 5 is a schematic diagram showing the configuration of an incineration facility according to the second embodiment. FIG. 6 is a block diagram showing the configuration of an information processing device according to the second embodiment. FIG. 7 is a flowchart illustrating an example of an information processing method by the information processing apparatus according to the second embodiment. FIG. 8 is a schematic diagram showing the configuration of an incineration facility according to the third embodiment. FIG. 9 is a block diagram showing the configuration of an information processing device according to the third embodiment. FIG. 10 is a flowchart illustrating an example of an information processing method by the information processing apparatus according to the third embodiment.

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Note that in the following explanation and the drawings used in the following explanation, the same reference numerals are used for parts that can be configured the same, and overlapping explanations are omitted.

As used herein, "A and/or B" means one or both of A and B. Furthermore, in this specification, "visible light" refers to electromagnetic waves with wavelengths of 360 nm to 830 nm, and "infrared rays" refers to electromagnetic waves with wavelengths of 830 nm to 14,000 nm.

(First embodiment)
FIG. 1 is a schematic diagram showing the configuration of an incineration facility 1 according to the first embodiment.

As shown in FIG. 1, the incineration facility 1 includes a platform 21 on which a transport vehicle (garbage collection truck) 22 loaded with waste stops, a garbage pit 3 in which the waste input from the platform 21 is stored, and a garbage pit 3 where the waste is stored. A crane 5 that stirs and transports the waste stored in the pit 3, a hopper 4 into which the waste transported by the crane 5 is thrown, and an incinerator 6 that burns the waste thrown from the hopper 4. The incinerator 6 includes an exhaust heat boiler 2 that recovers exhaust heat from exhaust gas generated within the incinerator 6. The type of incinerator 6 is not limited to a stoker furnace as shown in FIG. 1, but also includes a fluidized bed furnace (also referred to as a fluidized bed furnace). Further, the structure of the garbage pit 3 is not limited to the one-stage pit as shown in FIG. 1, but also includes a two-stage pit. The incineration facility 1 is also provided with a crane control device 50 that controls the operation of the crane 5 and a combustion control device 20 that controls the combustion of waste (combustible materials) in the incinerator 6.

The waste loaded on the transport vehicle 22 is introduced into the garbage pit 3 from the platform 21 and stored in the garbage pit 3. The waste stored in the garbage pit 3 is stirred by the crane 5 and transported to the hopper 4 by the crane 5, and is thrown into the incinerator 6 via the hopper 4. Disposed of by incineration.

As shown in FIG. 1, the incineration facility 1 according to the present embodiment includes an infrared imaging device 71 and a visible light imaging device 72 that image the inside of the incinerator 6, and information processing that determines the combustion state in the incinerator 6. A device 10 is provided.

In the illustrated example, the infrared imaging device 71 is installed above the downstream side of the incinerator 6. A plurality of infrared imaging devices 71 may be installed above the downstream side of the incinerator 6. The installation position of the infrared imaging device 71 is not limited to the upper part of the downstream side of the incinerator 6, and the infrared imaging device 71 can be installed horizontally on the downstream side of the incinerator 6, or above or horizontally on the upstream side of the incinerator 6. may be installed. The infrared imaging device 71 has an imaging wavelength (for example, 3.7 μm to 4.0 μm) appropriately selected in advance so as to cut the flame wavelength (for example, 4.1 μm to 4.5 μm), and the infrared imaging device 71 captures infrared imaging data (video As image data), it is possible to obtain an image including images of the actually burning materials from the upstream side of the incinerator 6, from which the influence of flame has been removed. The infrared imaging device 71 has an appropriate imaging wavelength (for example, 3.7 μm to 4.0 μm) in advance so as to further cut off the wavelength of CO ₂ and water vapor (for example, 4.0 μm to 4.2 μm) in addition to the flame wavelength. The infrared imaging data (video image data) includes an image from the upstream side of the incinerator 6 to the image of the actual burning material, with the effects of flame, CO ₂ and water vapor removed. It may be possible to obtain it. The frame rate of the infrared imaging device 71 does not need to be a particularly high frame rate, and may be a general frame rate (about 30 fps) or a low frame rate (about 5 to 10 fps).

In the illustrated example, the visible light imaging device 72 is installed above the downstream side of the incinerator 6, and visible light imaging data (moving image data) of the flame state (combustion state) of the material to be burned on the grate. ) can now be obtained. A plurality of visible light imaging devices 72 may be installed above the downstream side of the incinerator 6. Note that the installation position of the visible light imaging device 72 is not limited to the upper part on the downstream side of the incinerator 6, and the visible light imaging device 72 can be installed horizontally on the downstream side of the incinerator 6, or above or horizontally on the upstream side of the incinerator 6. It may be installed in the direction. The visible light imaging device 72 may be an RGB camera, a 3D camera, an RGB-D camera, or a combination of two or more of these. The frame rate of the visible light imaging device 72 does not need to be a particularly high frame rate, and may be a general frame rate (about 30 fps) or a low frame rate (about 5 to 10 fps). .

Next, the configuration of the information processing device 10 according to this embodiment will be described. FIG. 2 is a block diagram showing the configuration of the information processing device 10. The information processing device 10 is configured by one or more computers.

As shown in FIG. 2, the information processing device 10 includes a communication section 11, a control section 12, and a storage section 13. Each part is communicably connected to each other via a bus or network.

Among these, the communication unit 11 is a communication interface between the information processing device 10 and each of the infrared imaging device 71, the visible light imaging device 72, the crane control device 50, and the combustion control device 20. The communication unit 11 transmits and receives information between the information processing device 10 and each of the infrared imaging device 71, the visible light imaging device 72, the crane control device 50, and the combustion control device 20.

The storage unit 13 is, for example, a nonvolatile data storage such as a hard disk or flash memory. The storage unit 13 stores various data handled by the control unit 12. The storage unit 12 also stores a first algorithm 13a1 used for machine learning by a first model construction unit 12c1, which will be described later, a second algorithm 13a2 used for machine learning by the first model construction unit 12c1, and a first imaging data acquisition unit. 12a1, the second image data 13b2 obtained by the second image data acquisition section 12a2, the first teacher data 13c1 generated by the first teacher data generation section 12b, and the second image data 13b1 obtained by the second image data acquisition section 12a1. Second teacher data 13c2 generated by the teacher data generator 12b and a combustion state determination map 13d used by the combustion state determiner 12e are stored. Details of each piece of information 13a1 to 13d will be described later.

The control unit 12 is a control means that performs various processes of the information processing device 10. As shown in FIG. 2, the control unit 11 includes a first imaging data acquisition unit 12a1, a second imaging data acquisition unit 12b1, a first teacher data generation unit 12b1, a second teacher data generation unit 12b2, and a first A model construction section 12c1, a second model construction section 12c2, a first image analysis section 12d1, a second image analysis section 12d2, a combustion state determination section 12e, an instruction section 12f, and a process data acquisition section 12g. have. Each of these units may be realized by a processor within the information processing device 10 executing a predetermined program, or may be implemented by hardware.

The first imaging data acquisition unit 12a1 acquires infrared imaging data inside the incinerator 6 captured by the infrared imaging device 71 as first imaging data 13b1. The first imaging data 13b1 is stored in the storage unit 13.

The second imaging data acquisition unit 12a2 acquires visible light imaging data inside the incinerator 6 captured by the visible light imaging device 72 as second imaging data 13b2. The second image data 13b2 is stored in the storage unit 13.

The process data acquisition unit 12g obtains process data (for example, furnace outlet temperature and evaporation amount) measured by various sensors (not shown) installed in the incineration facility 1, and/or calculations from the process data. Obtain the calculated data. The process data and/or calculation data may be stored in the storage unit 13. The calculated data includes the difference value between the measured value PV and the set value SV (the absolute value of the difference value may also be used), the moving average value, maximum value, and minimum value over a predetermined period (for example, 1 minute, 10 minutes, 20 minutes, 1 hour). , a median value, an integral value, a differential value, a standard deviation, and a difference value from a value before a predetermined time (which may be an absolute value of the difference value).

The first teacher data generation unit 12b1 generates a classification label in at least one first evaluation axis, which is an element for determining the combustion state, and which is artificially assigned by the operator based on empirical rules to the first imaged data 13b1. By linking, the first teacher data 13c1 is generated. The first teacher data generation unit 12b1 divides the first imaged data 13b1 into a plurality of blocks (for example, a drying region, a gasification combustion region, and a dry combustion region in the combustion process), and the operator The first teacher data 13c1 may be generated by linking a classification label in at least one first evaluation axis, which is an element for determining the combustion state, which is artificially assigned based on an empirical rule. The first teacher data 13c1 is stored in the storage unit 13. Here, the classification label may be absolute classification information on the first evaluation axis (classification determined by the relationship to a predetermined absolute standard (threshold value)), or may be information on the classification between multiple pieces of infrared imaging data. may be a relative order on the first evaluation axis (a relative order relationship within a group consisting of a plurality of infrared imaging data). The first evaluation axis is at least one of the following: the amount of materials to be burned, the quality of materials to be burned, the type of materials to be burned, the temperature of materials to be burned, the amount of combustion gas generated (CO, etc.), and the temperature of the wall of the incinerator. It may include one. The amount of the combustible material is a value corresponding to, for example, the volume, weight, density, cross-sectional area, etc. of the combustible material. The quality of the combustible material is a value corresponding to, for example, the combustibility, calorific value, moisture, density, etc. of the combustible material. The types of materials to be burnt include, for example, unbroken garbage bags, paper garbage, pruned branches, bedding, sludge, coarse crushed garbage, cardboard, jute bags, paper bags, and bottom garbage (located near the bottom of the garbage pit 3 and containing information). waste (compressed waste with high moisture content), wood chips, textile waste, clothing waste, plastic waste, animal residue, animal carcasses, kitchen waste, plants, soil, medical waste, incinerated ash, bicycles, chests of drawers, Beds, shelves, desks, chairs, agricultural vinyl, plastic bottles, Styrofoam, meat and bone meal, agricultural products, ceramics, glass scraps, metal scraps, rubble, concrete scraps, tatami mats, bamboo, straw, and activated carbon. The temperature of the object to be combusted is, for example, a value corresponding to the temperature of the object to be combusted. The amount of combustion gas generated is a value corresponding to the amount of gas generated during the combustion process, such as CO, hydrogen, hydrocarbons, NOx, SOx, HCl, and dioxins. The temperature of the wall of the incinerator is a value corresponding to, for example, the temperature of the wall and ceiling of the incinerator.

For example, if the first evaluation axis is the amount of combustible material, the classification label may be large, normal, or small (absolute information) (the types of labels may be increased), or there may be multiple classification labels. It may be information in which the magnitude of the amount of burnt material is ordered (relative information) with respect to the first image data of . As another example, if the quality of the material to be combusted is the evaluation axis, the classification label is normal garbage, high quality garbage (i.e. high calorie garbage), low quality garbage (i.e. low calorie garbage) (absolute information). (the number of types of labels may be increased), or it may be information in which the quality of the burnt material is ordered (relative information) with respect to a plurality of first imaging data. As another example, when the first evaluation axis is the type of combustible material, the classification label may be garbage in unbroken garbage bags, pruned branches, or futons (absolute information) (the types of labels may be increased). ). As another example, if the amount of combustion gas generated (CO, etc.) is the evaluation axis, the classification label may be large, normal, or small (absolute information) (the types of labels may be increased). , it may be information (relative information) in which the magnitudes of combustion gas generation amounts (CO, etc.) are ordered for a plurality of first imaging data. As another example, if the temperature of the wall of the incinerator is the first evaluation axis, the classification label may be high, normal, or low (the types of labels may be increased), and the classification label may be high, normal, or low (the types of labels may be increased), and the The data may be information in which the temperature of the wall of the incinerator is ordered (relative information).

The first teacher data generation unit 12b1 generates the first image data 13b1 and process data obtained from various sensors (not shown) installed in the incineration facility 1 and/or the amount of calculation obtained by calculation from the process data. The first teacher data 13c1 may be generated by associating a classification label in the first evaluation axis that is artificially assigned by an operator based on an empirical rule to the combination. Here, the process data is data that takes into account the time lag between the imaging time of the first imaging data and the response time of the sensor. The time lag may be determined as appropriate based on the response speed of the sensor, the installation position of the sensor within the facility, experiments, simulations, and the operator's empirical rules.

The second teacher data generation unit 12b2 includes a classification label in at least one second evaluation axis, which is an element for determining the combustion state, and which is artificially assigned by the operator based on empirical rules to the second imaged data 13b1. By linking, the second teacher data 13c2 is generated. The second teacher data generation unit 12b2 divides the second imaged data 13b1 into a plurality of blocks (for example, a drying region, a gasification combustion region, and a combustion region in the combustion process), and the operator The second teacher data 13c2 may be generated by linking a classification label in at least one first evaluation axis, which is an element for determining the combustion state, which is artificially assigned based on an empirical rule. The second teacher data 13c2 is stored in the storage unit 13. Here, the classification label may be absolute classification information on the second evaluation axis (classification determined by the relationship to a predetermined absolute standard (threshold value)), or may be between multiple pieces of visible light imaging data. It may be the relative order on the second evaluation axis (relative order relationship within a group consisting of a plurality of visible light imaging data). The second evaluation axis is the flame condition, the position of the combustion completion point (burnout point), the shape of the combustion completion point, the amount of combustible material, the quality of combustible material, the type of combustible material, the amount of unburned material, amount of incinerated ash. The flame state is a value corresponding to, for example, the strength of the flame, the brightness of the flame, and the like. The position of the combustion completion point is, for example, a value corresponding to the position on the grate that constitutes the incinerator. The shape of the combustion completion point is a value corresponding to, for example, the curvature and radius of curvature of the combustion completion point. The amount of the combustible material is a value corresponding to, for example, the volume, weight, density, cross-sectional area, etc. of the combustible material. The quality of the combustible material is a value corresponding to, for example, the combustibility, calorific value, moisture, density, etc. of the combustible material. The types of materials to be burnt include, for example, unbroken garbage bags, paper garbage, pruned branches, bedding, sludge, coarse crushed garbage, cardboard, jute bags, paper bags, and bottom garbage (located near the bottom of the garbage pit 3 and containing information). waste (compressed waste with high moisture content), wood chips, textile waste, clothing waste, plastic waste, animal residue, animal carcasses, kitchen waste, plants, soil, medical waste, incinerated ash, bicycles, chests of drawers, Beds, shelves, desks, chairs, agricultural vinyl, plastic bottles, Styrofoam, meat and bone meal, agricultural products, ceramics, glass scraps, metal scraps, rubble, concrete scraps, tatami mats, bamboo, straw, and activated carbon. The amount of unburned material is a value corresponding to, for example, the volume, weight, density, cross-sectional area, amount, etc. of unburned material. The amount of incinerated ash is a value corresponding to, for example, the volume, weight, density, cross-sectional area, etc. of incinerated ash.

As an example, when the second evaluation axis is the flame condition, the classification label may be good, normal, or bad (absolute information) (the number of label types may be increased), or it may be multiple (for example, two It may also be information in which the superiority or inferiority of the flame state is ordered (relative information) with respect to the second imaged data of (1). As another example, if the position of the combustion completion point (burnout point) is used as the evaluation axis, the classification label may be near, slightly near, normal, slightly back, or far (the types of labels may be increased). ) However, the distance of the position of the combustion completion point may be ordered for the plurality of second imaging data. As another example, if the shape of the combustion completion point is used as the evaluation axis, the classification label may be good, normal, or bad (the types of labels may be increased), and the classification label may be good, normal, or bad (the types of labels may be increased), and the classification label may be , the shape of the combustion completion point may be ranked in order. As another example, when the second evaluation axis is the amount of combustible material, the classification label may be large, normal, or small (the types of labels may be increased), and the classification label may be large, normal, or small (the types of labels may be increased), and the It may also be an arrangement in which the amount of combustible material is ordered in order of magnitude. As another example, when the second evaluation axis is the quality of the material to be combusted, the classification label may be ordinary garbage, high quality garbage (i.e. high calorie garbage), or low quality garbage (i.e. low calorie garbage) (label (the number of types may be increased), or the quality of the burnt material may be ranked in order of quality for the plurality of second imaging data. As another example, when the second evaluation axis is the type of material to be burned, the classification label may be garbage bag-unbroken garbage, pruned branches, or futon (the number of types of labels may be increased). As another example, when the second evaluation axis is the amount of unburned materials, the classification label may be large, normal, or small (the types of labels may be increased), and the classification label may be large, normal, or small (the types of labels may be increased), and the The amount of unburned material may be ordered according to the amount of unburned material. As another example, when the amount of incinerated ash is used as the second evaluation axis, the classification label may be large, normal, or small (the types of labels may be increased), and the classification label may be “large”, “normal”, or “small” (the types of labels may be increased). On the other hand, the amount of incinerated ash may be ordered in order of magnitude.

The second teacher data generation unit 12b2 generates a combination of the second imaging data 13b2 and process data obtained from various sensors (not shown) installed in the incineration facility 1 and/or the amount of calculation obtained by calculation from the process data. The second teacher data 13c2 may be generated by associating a classification label in the second evaluation axis that is artificially assigned by the operator based on empirical rules to the combination. Here, the process data is data that takes into account the time lag between the imaging time of the second imaging data and the response time of the sensor. The time lag may be determined as appropriate based on the response speed of the sensor, the installation position of the sensor within the facility, experiments, simulations, and the operator's empirical rules.

The first imaging data 13b1 and/or the second imaging data 13b2 may be moving image data within 60 seconds, or may be moving image data within 30 seconds. According to the findings of the present inventors, the combustion state within the incinerator 6 changes every 5 to 10 seconds. Therefore, by using moving image data within 60 seconds (or within 30 seconds) as one segment as the first imaging data 13b1 and/or the second imaging data 13b2, the combustion state in the incinerator 6 can be determined more accurately. It becomes possible to classify, predict, and estimate.

Further, the first image data 13b1 and/or the second image data 13b2 may be moving image data of 5 seconds or more, or may be 7 seconds or more. While the combustion state in the incinerator 6 changes every 5 to 10 seconds, for example, moving image data of 5 seconds or more (or 7 seconds or more) is used as the first image data 13b1 and/or the second image data 13b2. By using it as a delimiter, it becomes possible to classify, predict, and estimate the combustion state in the incinerator 6 with higher accuracy.

The first model construction unit 13c1 generates a first learned model by performing machine learning on the first teacher data 13b1 generated by the first teacher data generation unit 12b1 using the first algorithm 13a1. The first algorithm 13a1 used for machine learning includes statistical estimation using maximum likelihood classification method, Boltzmann machine, neural network (NN), support vector machine (SVM), Bayesian network, sparse regression, decision tree, random forest, It may include at least one of boosting, reinforcement learning, and deep learning.

Similarly, the second model construction unit 13c2 generates a second learned model by performing machine learning on the second teacher data 13b2 generated by the second teacher data generation unit 12b2 using the second algorithm 13a2. . The second algorithm 13a2 used for machine learning includes statistical estimation using maximum likelihood classification, Boltzmann machine, neural network (NN), support vector machine (SVM), Bayesian network, sparse regression, decision tree, and random forest. It may include at least one of boosting, reinforcement learning, and deep learning.

The first image analysis unit 13d1 inputs the new first imaging data (infrared imaging data) inside the incinerator 6 acquired by the first imaging data acquisition unit 13a1, and receives the new first imaging data (infrared imaging data) generated by the first model construction unit 13c1. Using the first learned model, the evaluation result on the first evaluation axis (for example, the amount of combustible material) is obtained as output data. The first image analysis unit 13d1 may normalize (score) the evaluation result on the first evaluation axis into a numerical range of 0 to 100, for example, and obtain the numerical value (score) within the range as output data. .

The second image analysis unit 13d2 inputs the new second image data (visible light image data) inside the incinerator 6 acquired by the second image data acquisition unit 13a2, and generates the image data generated by the second model construction unit 13c2. Using the second learned model, the evaluation result on the second evaluation axis (for example, flame state) is obtained as output data. The second image analysis unit 13d2 may normalize (score) the evaluation result on the second evaluation axis into a numerical range of 0 to 100, for example, and obtain the numerical value (score) within the range as output data. .

The combustion state determination unit 12e determines whether the incinerator is in Determine the current combustion state within 6. The combustion status determination result may be a label indicating the characteristics of the combustion status (over-combustion, thick garbage layer combustion, thin garbage layer combustion, garbage withered garbage, low-quality garbage combustion, etc.), or a numerically modified version of the label. It may be. Over-combustion refers to a combustion state in which combustion is active, for example, when a large amount of high-quality (easily combustible) garbage is present in the material to be burned, and the combustion temperature becomes extremely high. Thick garbage combustion is, for example, a combustion state in which the amount of materials to be combusted is larger than normal. The garbage layer thin combustion is, for example, a combustion state in which the amount of materials to be burned is smaller than normal. Garbage drying is, for example, a combustion state in which the amount of materials to be combusted is extremely smaller than normal, and the combustion temperature is low. Low-quality waste combustion refers to a combustion state in which combustion is weak, such as when a large amount of low-quality (low-calorie waste, etc.) waste is present among the objects to be burned, and the combustion temperature is low.

The combustion state determination unit 12e calculates a set (X, Y) consisting of the evaluation result (X) on the first evaluation axis and the evaluation result (Y) on the second evaluation axis, with the first evaluation axis being the X coordinate axis and the second evaluation axis. The current combustion state in the incinerator 6 may be determined by mapping onto a predetermined combustion state determination map 13d whose evaluation axis is the Y coordinate axis. The combustion state determination unit 12e selects, for example, a set (X, Y, Z, . . . ) of evaluation results on one or more first evaluation axes and evaluation results on one or more second evaluation axes. A predetermined N-dimensional space where one of the evaluation axes is the X coordinate axis, another of the first evaluation axes is the Y coordinate axis, and one of the second evaluation axes is the Z coordinate axis (N is 3 or more The current combustion state in the incinerator 6 may be determined by mapping it onto the combustion state determination map 13d.

FIG. 4 is a diagram showing an example of the combustion state determination map 13d. In the combustion state determination map 13d shown in FIG. It is divided into six areas: thin combustion zone, good combustion (normal) zone, thick garbage combustion zone, and low-quality garbage combustion zone. The combustion state determination unit 12e determines the evaluation result (X) on the first evaluation axis and the evaluation result (Y) on the second evaluation axis when the amount of combustible material is the first evaluation axis and the flame state is the second evaluation axis. (X, Y) is mapped onto this combustion state determination map 13d. For example, as shown in FIG. 4, a point P indicating a set (X, Y) consisting of the evaluation result (X) on the first evaluation axis and the evaluation result (Y) on the second evaluation axis indicates good combustion (normal). When mapped within the zone, the combustion state determining unit 12e determines that the current combustion state is "good combustion (normal)". The range of the above region (threshold value (position, length and shape of boundary line)) may be determined from the operator's empirical rules, or may be determined based on the evaluation result (X) on the first evaluation axis and the second evaluation axis. The correspondence between the evaluation result (Y) and the process data is calculated using maximum likelihood classification, Boltzmann machine, neural network (NN), support vector machine (SVM), Bayesian network, sparse regression, decision tree, and random forest. It may be determined by machine learning using an algorithm including at least one of statistical estimation, Kalman filter, autoregression, boosting, reinforcement learning, and deep learning.

The combustion state determination unit 12e corrects the combustion state determination result according to process data obtained from various sensors (not shown) installed in the facility and/or the amount of calculation obtained by calculation from the process data. Good too. For example, if the combustion state determination result is "low-quality waste combustion" and the furnace outlet temperature or evaporation amount difference value PV-SV satisfies a predetermined condition, the combustion state determination result is It may be corrected to "good combustion (normal)".

Depending on the combustion state determination result, the combustion state determination unit 12e may display an alert such as that it is difficult to maintain the performance management index value within a predetermined range on a display (not shown), or may display an alert using audio or vibration. , the alarm may be issued by means of light, etc.

The instruction section 12f transmits an operation instruction to the crane control device 50 and/or the combustion control device 20 based on the determination result of the combustion state determination section 12e. For example, the instruction section 12f transmits an operation instruction predefined for each zone on the combustion state determination map to the crane control device 50 and/or the combustion control device 20 based on the determination result of the combustion state determination section 12e. . Specifically, for example, the instruction unit 12f issues operation instructions suitable for the current combustion state (increase or decrease the combustion temperature, increase or decrease the combustion time, increase or decrease the amount of air, increase or decrease the amount of waste to be fed, etc.). decrease) is sent to the combustion control device 20. As an example, when the current combustion state is "overcombustion", the instruction unit 12f may transmit an operation instruction such as a reduction in the amount of air or a reduction in the amount of garbage sent to the combustion control device 20. As another example, when the current combustion state is "thick garbage layer combustion," the instruction unit 12f issues operational instructions to the combustion control device 20, such as reducing the amount of garbage fed, in order to thin the layer of garbage. You can also send it. As another example, when the current combustion state is "thin combustion", the instruction unit 12f may issue operational instructions such as reducing the amount of air or reducing the amount of garbage fed in order to thicken the garbage layer. may be transmitted to the combustion control device 20. As another example, when the current combustion state is "garbage dry", the instruction unit 12f may take measures such as decreasing the amount of air or increasing the amount of garbage fed in order to increase the amount of garbage into the incinerator. The operation instruction may be transmitted to the combustion control device 20. As another example, when the current combustion state is "low-quality waste combustion", the instruction unit 12f provides operational instructions such as increasing the amount of air and decreasing the amount of waste fed in order to properly burn the existing waste. may be transmitted to the combustion control device 20.

Next, an example of an information processing method by the information processing apparatus 10 having such a configuration will be described. FIG. 3 is a flowchart illustrating an example of an information processing method.

As shown in FIG. 3, first, before the incineration facility 1 is operated, the first teacher data generation unit 12b1 generates past infrared imaging data (first imaging data 13b1) inside the incinerator 6 captured by the infrared imaging device 71. ), the first training data 13c1 is generated by linking a classification label in at least one first evaluation axis, which is an element for determining the combustion state, which is artificially assigned by the operator based on empirical rules. do. In addition, the second teacher data generation unit 12b2 generates an image of past visible light imaging data (second imaging data 13b2) inside the incinerator 6 captured by the visible light imaging device 72, based on the operator's empirical rules. The second teacher data 13c2 is generated by linking the classification labels in at least one second evaluation axis that are assigned as elements for determining the combustion state (step S11).

Next, the first model construction unit 13c1 generates a first learned model by performing machine learning on the first teacher data 13b1 generated by the first teacher data generation unit 12b1 using the first algorithm 13a1. . Further, the second model construction unit 13c2 generates a second learned model by performing machine learning on the second teacher data 13b2 generated by the second teacher data generation unit 12b2 using the second algorithm 13a2 ( Step S12).

Next, while the incineration facility 1 is operating, the first imaging data acquisition unit 12a1 acquires new infrared imaging data inside the incinerator 6 captured by the infrared imaging device 71 as the first imaging data 13b1, and also acquires new infrared imaging data as the first imaging data 13b1. The imaging data acquisition unit 12a2 acquires visible light imaging data inside the incinerator 6 captured by the visible light imaging device 72 as second imaging data 13b2 (step S13).

Next, the first image analysis section 13d1 inputs the new first imaging data (infrared imaging data) in the incinerator 6 acquired by the first imaging data acquisition section 13a1, and generates a model by the first model construction section 13c1. Using the first learned model, the evaluation result on the first evaluation axis (for example, the amount of combustible material) is obtained as output data. In addition, the second image analysis unit 13d2 receives new second image data (visible light image data) in the incinerator 6 acquired by the second image data acquisition unit 13a2, and generates data using the second model construction unit 13c2. Using the second learned model, the evaluation result on the second evaluation axis (for example, flame state) is obtained as output data (step S14).

Then, the combustion state determination section 12e determines the condition of the inside of the incinerator 6 based on the evaluation result on the first evaluation axis by the first image analysis section 13d1 and the evaluation result on the second evaluation axis by the second image analysis section 13d2. The current combustion state is determined (step 15).

In step S15, the combustion state determination unit 12e calculates the set (X, Y) consisting of the evaluation result (X) on the first evaluation axis and the evaluation result (Y) on the second evaluation axis, with the first evaluation axis being The current combustion state in the incinerator 6 may be determined by mapping onto a predetermined combustion state determination map 13d in which the Y coordinate axis is the coordinate axis and the second evaluation axis (see FIG. 4).

Thereafter, the instruction section 12f transmits an operation instruction to the crane control device 50 and/or the combustion control device 20 based on the determination result of the combustion state determination section 12e (step S16).

According to the present embodiment as described above, the infrared imaging device 71 removes the influence of flame as the infrared imaging data (first imaging data) inside the incinerator 6 by appropriately selecting the imaging wavelength. Since it is possible to obtain an image including images of the actually burning materials from the upstream side of the incinerator 6, infrared imaging data inside the incinerator 6 can be used to at least provide an element for determining the combustion state. By performing an evaluation on one first evaluation axis and determining the combustion state based on the evaluation results, it is possible to classify, predict, and estimate the combustion state and predict and process values from only the combustion image that can only grasp the flame situation. Compared to the case where estimation is performed, the combustion state in the incinerator 6 can be classified, predicted, and estimated with higher accuracy.

Further, according to the present embodiment, by determining the combustion state based on the evaluation results on each of the first evaluation axis X and the second evaluation axis Y, the combustion state is determined based on the evaluation result on one evaluation axis. Compared to the case where the state is determined, the combustion state in the incinerator 6 can be classified, predicted, and estimated with higher accuracy. In addition, by determining the combustion state by mapping the evaluation results on each of the first evaluation axis X and the second evaluation axis Y onto the combustion state determination map 13d, the first evaluation axis X and the second evaluation axis Y It is possible to intuitively understand the relationship between the evaluation results for each of the above and the combustion state as a determination result, and it is also possible to quickly determine the combustion state.

(Second embodiment)
Next, a second embodiment will be described with reference to FIGS. 5 to 7. FIG. 5 is a schematic diagram showing the configuration of an incineration facility 1 according to the second embodiment.

As shown in FIG. 5, the incineration facility 1 according to the second embodiment differs from the first embodiment only in that the visible light imaging device 72 that images the inside of the incinerator 6 is omitted, and other points The configuration is similar to that of the first embodiment.

FIG. 6 is a block diagram showing the configuration of the information processing device 10 according to the second embodiment. The information processing device 10 is configured by one or more computers.

As shown in FIG. 6, the information processing device 10 includes a communication section 11, a control section 12, and a storage section 13. Each part is communicably connected to each other via a bus or network.

Among these, the communication unit 11 is a communication interface between the information processing device 10 and the infrared imaging device 71, the crane control device 50, and the combustion control device 20. The communication unit 11 transmits and receives information between the infrared imaging device 71, the crane control device 50, and the combustion control device 20, and the information processing device 10.

The storage unit 13 is, for example, a nonvolatile data storage such as a hard disk or flash memory. The storage unit 13 stores various data handled by the control unit 12. The storage unit 12 also stores a first algorithm 13a1 used for machine learning by a first model construction unit 12c1 (described later), first imaging data 13b1 acquired by a first imaging data acquisition unit 12a1, and first training data generation. First teacher data 13c1 generated by the section 12b1 and a combustion state determination map 13d used by the combustion state determination section 12e are stored. Details of each piece of information 13a1 to 13d will be described later.

The control unit 12 is a control means that performs various processes of the information processing device 10. As shown in FIG. 6, the control unit 11 includes a first imaging data acquisition unit 12a1, a first teacher data generation unit 12b1, a first model construction unit 12c1, a first image analysis unit 12d1, and a combustion state determination unit. 12e, an instruction section 12f, and a process data acquisition section 12g. Each of these units may be realized by a processor within the information processing device 10 executing a predetermined program, or may be implemented by hardware.

Among these, the configurations of the first imaging data acquisition section 12a1, the process data acquisition section 12g, and the instruction section 12f are the same as those in the first embodiment, and the description thereof will be omitted.

The first teacher data generation unit 12b1 generates each of two or more first evaluation axes, which are factors for determining the combustion state, that are artificially assigned to the first imaging data 13b1 by an operator based on empirical rules. The first teacher data 13c1 is generated by linking the classification labels in . The first teacher data 13c1 is stored in the storage unit 13. Here, the classification label may be absolute classification information on the first evaluation axis (classification determined by the relationship to a predetermined absolute standard (threshold value)), or may be information on the classification between multiple pieces of infrared imaging data. may be a relative order on the first evaluation axis (a relative order relationship within a group consisting of a plurality of infrared imaging data). The first evaluation axis is the flame condition, the amount of materials to be burned, the quality of materials to be burned, the type of materials to be burned, the temperature of materials to be burned, the amount of combustion gas generated (CO, etc.), and the temperature of the wall of the incinerator. It may include two or more of the following. The flame state can be determined from the amount of gaseous fluctuations reflected in the first image data 13b1 (infrared image data).

As an example, if the flame condition is the first evaluation axis, the classification label may be good, normal, or bad (absolute information) (the number of types of labels may be increased), or it may include multiple classification labels. It may be information in which the superiority or inferiority of the flame state is ordered (relative information) with respect to the first imaged data. As another example, if the amount of combustibles is the first evaluation axis, the classification label may be large, normal, or small (absolute information) (the types of labels may be increased) However, the plurality of first imaging data may be ordered in terms of the amount of burnt material (relative information). As another example, if the quality of the material to be combusted is one of the first evaluation axis, the classification labels are normal garbage, high quality garbage (i.e. high calorie garbage), low quality garbage (i.e. low calorie garbage) (absolute information). ) (the types of labels may be increased), or may be information in which the quality of the burnt material is ordered (relative information) with respect to a plurality of first imaging data. As another example, if the type of material to be burned is one of the first evaluation axes, the classification label may be garbage bag unbroken garbage, pruned branches, futon (absolute information) (label type may be increased). As another example, if the amount of combustion gas generated (CO, etc.) is the first evaluation axis, the classification label may be large, normal, or small (absolute information) (the types of labels may be increased). Alternatively, the information may be information (relative information) in which the magnitude of the amount of combustion gas generated (CO, etc.) is ordered for the plurality of first imaging data. As another example, if the temperature of the wall of the incinerator is used as one first evaluation axis, the classification label may be high, normal, or low (the types of labels may be increased), or there may be multiple classification labels. It may be information in which the temperature of the wall of the incinerator is ordered (relative information) with respect to one piece of imaging data.

The first teacher data generation unit 12b1 generates the first image data 13b1 and process data obtained from various sensors (not shown) installed in the incineration facility 1 and/or the amount of calculation obtained by calculation from the process data. The first teacher data 13c1 may be generated by associating classification labels in each of two or more first evaluation axes artificially assigned by an operator based on empirical rules to the combination. Here, the process data is data that takes into account the time lag between the imaging time of the first imaging data and the response time of the sensor. The time lag may be determined as appropriate based on the response speed of the sensor, the installation position of the sensor within the facility, experiments, simulations, and the operator's empirical rules.

The first imaging data 13b1 may be moving image data within 60 seconds, or may be moving image data within 30 seconds. According to the findings of the present inventors, the combustion state within the incinerator 6 changes every 5 to 10 seconds. Therefore, by using moving image data within 60 seconds (or within 30 seconds) as one segment as the first imaged data 13b1, the combustion state in the incinerator 6 can be classified, predicted, and estimated with higher accuracy. It becomes possible.

Further, the first imaging data 13b1 may be moving image data of 5 seconds or more, or may be 7 seconds or more. While the combustion state in the incinerator 6 changes every 5 to 10 seconds, by using moving image data of 5 seconds or more (or 7 seconds or more) as the first image data 13b1, for example, the incinerator 6 changes every 5 to 10 seconds. It becomes possible to classify, predict, and estimate the combustion state in the engine 6 with higher accuracy.

The first model construction unit 13c1 generates a first learned model by performing machine learning on the first teacher data 13b1 generated by the first teacher data generation unit 12b1 using the first algorithm 13a1. The first algorithm 13a1 used for machine learning includes statistical estimation using maximum likelihood classification method, Boltzmann machine, neural network (NN), support vector machine (SVM), Bayesian network, sparse regression, decision tree, random forest, It may include at least one of boosting, reinforcement learning, and deep learning.

The first image analysis unit 13d1 inputs the new first imaging data (infrared imaging data) inside the incinerator 6 acquired by the first imaging data acquisition unit 13a1, and receives the new first imaging data (infrared imaging data) generated by the first model construction unit 13c1. Using the first learned model, evaluation results on each of two or more first evaluation axes (for example, flame condition and amount of combustible material) are obtained as output data. The first image analysis unit 13d1 normalizes (scores) the evaluation results on each of the two or more first evaluation axes into a numerical range of 0 to 100, for example, and converts the numerical values (scores) within the range into output data. You can also obtain it as

The combustion state determination unit 12e determines the current combustion state in the incinerator 6 based on the evaluation results (scores) on each of the two or more first evaluation axes by the first image analysis unit 13d1. The combustion status determination result may be a label indicating the characteristics of the combustion status (over-combustion, thick garbage layer combustion, thin garbage layer combustion, garbage withered garbage, low-quality garbage combustion, etc.), or a numerically modified version of the label. It may be.

The combustion state determination unit 12e places the set of evaluation results on each of the two or more first evaluation axes onto a predetermined combustion state determination map 13d having the two or more first evaluation axes as coordinate axes. The current combustion state within the incinerator 6 may be determined by mapping.

FIG. 4 is a diagram showing an example of the combustion state determination map 13d. In the combustion state determination map 13d shown in FIG. It is divided into six areas: thin combustion zone, good combustion (normal) zone, thick garbage combustion zone, and low-quality garbage combustion zone. The combustion state determination unit 12e uses the set (X, Y) consisting of evaluation results on each first evaluation axis when the amount of combustible material and the flame state are two different first evaluation axes to determine the combustion state. It is mapped on the map 13d. For example, as shown in FIG. 4, a set (X, Y ) is mapped within the good combustion (normal) zone, the combustion state determination unit 12e determines that the current combustion state is "good combustion (normal)."

The combustion state determination unit 12e corrects the combustion state determination result according to process data obtained from various sensors (not shown) installed in the facility and/or the amount of calculation obtained by calculation from the process data. Good too. For example, if the combustion state determination result is "low-quality waste combustion" and the furnace outlet temperature or evaporation amount difference value PV-SV satisfies a predetermined condition, the combustion state determination result is It may be corrected to "good combustion (normal)".

Depending on the combustion state determination result, the combustion state determination unit 12e may display an alert such as that it is difficult to maintain the performance management index value within a predetermined range on a display (not shown), or may display an alert using audio or vibration. , the alarm may be issued by means of light, etc.

Next, an example of an information processing method by the information processing apparatus 10 having such a configuration will be described. FIG. 7 is a flowchart illustrating an example of an information processing method.

As shown in FIG. 7, first, before the incineration facility 1 is operated, the first teacher data generation unit 12b1 generates past infrared imaging data (first imaging data 13b1) inside the incinerator 6 captured by the infrared imaging device 71. ), the first training data 13c1 is created by linking two or more classification labels in the first evaluation axis, which are factors for determining the combustion state, which are artificially assigned by the operator based on empirical rules. is generated (step S21).

Next, the first model construction unit 13c1 generates a first learned model by performing machine learning on the first teacher data 13b1 generated by the first teacher data generation unit 12b1 using the first algorithm 13a1. (Step S22).

Next, while the incineration facility 1 is in operation, the first imaging data acquisition unit 12a1 acquires new infrared imaging data inside the incinerator 6 captured by the infrared imaging device 71 as the first imaging data 13b1 (step S23). .

Next, the first image analysis section 13d1 inputs the new first imaging data (infrared imaging data) in the incinerator 6 acquired by the first imaging data acquisition section 13a1, and generates a model by the first model construction section 13c1. Using the first learned model, evaluation results on each of two or more first evaluation axes (for example, amount of combustible material and flame condition) are obtained as output data (step S24).

Then, the combustion state determination unit 12e determines the current combustion state in the incinerator 6 based on the evaluation result on the first evaluation axis by the first image analysis unit 13d1 (step 25).

In step S25, the combustion state determination unit 12e converts a set of evaluation results on each of the two or more first evaluation axes into a predetermined combustion state determination map having the two or more first evaluation axes as coordinate axes. By mapping onto 13d, the current combustion state within the incinerator 6 may be determined (see FIG. 4).

Thereafter, the instruction section 12f transmits an operation instruction to the crane control device 50 and/or the combustion control device 20 based on the determination result of the combustion state determination section 12e (step S16).

According to the present embodiment as described above, the infrared imaging device 71 removes the influence of flame as the infrared imaging data (first imaging data) inside the incinerator 6 by appropriately selecting the imaging wavelength. Since it is possible to acquire an image including images of the actually burning materials from the upstream side of the incinerator 6, infrared imaging data inside the incinerator 6 can be used to determine the combustion state. By performing evaluation on more than one first evaluation axis and determining the combustion state based on the evaluation results, classification, prediction, and estimation of the combustion state and prediction of process values can be made from only combustion images that can only grasp the flame situation. - Compared to the case where estimation is performed, the combustion state in the incinerator 6 can be classified, predicted, and estimated with higher accuracy.

Further, according to the present embodiment, by determining the combustion state based on the evaluation results on each of the two or more first evaluation axes, the combustion state is determined based on the evaluation result on one evaluation axis. The combustion state in the incinerator 6 can be classified, predicted, and estimated with higher accuracy than in the case of Furthermore, by determining the combustion state by mapping the evaluation results for each of the two or more first evaluation axes onto the combustion state determination map 13d, the evaluation results for each of the two or more first evaluation axes can be determined. It is possible to intuitively grasp the relationship between the combustion state and the combustion state that is the determination result, and it is also possible to quickly determine the combustion state.

(Third embodiment)
Next, a third embodiment will be described with reference to FIGS. 8 to 10. FIG. 8 is a schematic diagram showing the configuration of an incineration facility 1 according to the third embodiment.

As shown in FIG. 8, the incineration facility 1 according to the third embodiment differs from the first embodiment only in that the infrared imaging device 71 that images the inside of the incinerator 6 is omitted, and other The configuration is similar to the first embodiment.

FIG. 9 is a block diagram showing the configuration of an information processing device 10 according to the third embodiment. The information processing device 10 is configured by one or more computers.

As shown in FIG. 9, the information processing device 10 includes a communication section 11, a control section 12, and a storage section 13. Each part is communicably connected to each other via a bus or network.

Among these, the communication unit 11 is a communication interface between the visible light imaging device 72, the crane control device 50, and the combustion control device 20, and the information processing device 10. The communication unit 11 transmits and receives information between the visible light imaging device 72, the crane control device 50, and the combustion control device 20, and the information processing device 10.

The storage unit 13 is, for example, a nonvolatile data storage such as a hard disk or flash memory. The storage unit 13 stores various data handled by the control unit 12. The storage unit 12 also stores a second algorithm 13a2 used for machine learning by a second model construction unit 12c2, which will be described later, second imaging data 13b2 acquired by a second imaging data acquisition unit 12a2, and second training data generation. Second teacher data 13c2 generated by the section 12b2 and a combustion state determination map 13d used by the combustion state determination section 12e are stored. Details of each piece of information 13a2 to 13d will be described later.

The control unit 12 is a control means that performs various processes of the information processing device 10. As shown in FIG. 9, the control unit 11 includes a second imaging data acquisition unit 12a2, a second teacher data generation unit 12b2, a second model construction unit 12c2, a second image analysis unit 12d2, and a combustion state determination unit. 12e, an instruction section 12f, and a process data acquisition section 12g. Each of these units may be realized by a processor within the information processing device 10 executing a predetermined program, or may be implemented by hardware.

Among these, the configurations of the second imaging data acquisition section 12a2, the process data acquisition section 12g, and the instruction section 12f are the same as those in the first embodiment, and a description thereof will be omitted.

The second teacher data generation unit 12b2 generates each of two or more second evaluation axes, which are factors for determining the combustion state, that are artificially assigned to the second imaging data 13b2 by an operator based on empirical rules. By linking the classification labels in , second teacher data 13c2 is generated. The second teacher data 13c2 is stored in the storage unit 13. Here, the classification label may be absolute classification information on the second evaluation axis (classification determined by the relationship to a predetermined absolute standard (threshold value)), or may be between multiple pieces of visible light imaging data. It may be the relative order on the second evaluation axis (relative order relationship within a group consisting of a plurality of visible light imaging data). The second evaluation axis is the flame condition, the position of the combustion completion point, the shape of the combustion completion point, the amount of combustible material, the quality of the combustible material, the type of combustible material, the amount of unburned material, and the amount of incinerated ash. It may include two or more of these.

As an example, if the flame condition is one second evaluation axis, the classification label may be good, normal, or bad (absolute information) (the types of labels may be increased), or there may be multiple classification labels. It may be information in which the superiority or inferiority of the flame state is ordered with respect to the second imaged data (relative information). As another example, if the position of the combustion completion point (burnout point) is one of the second evaluation axes, the classification label may be near, slightly near, normal, slightly back, or far (the type of label is (may be increased), or the distance of the position of the combustion completion point may be ordered with respect to the plurality of second imaging data. As another example, if the shape of the combustion completion point is used as one second evaluation axis, the classification label may be good, normal, or bad (the types of labels may be increased), and the classification label may be good, normal, or bad (the types of labels may be increased), and the The data may be ordered based on the shape of the combustion completion point. As another example, if the amount of combustible material is one of the second evaluation axis, the classification label may be large, normal, or small (absolute information) (the types of labels may be increased) However, the plurality of second imaging data may be ordered in terms of the amount of the burnt material (relative information). As another example, if the quality of the material to be combusted is one of the second evaluation axis, the classification label may be normal garbage, high quality garbage (i.e. high calorie garbage), or low quality garbage (i.e. low calorie garbage). (The number of types of labels may be increased.) Alternatively, the plurality of second image data may be ranked in terms of quality of the burnable material. As another example, when the type of burnt material is one of the second evaluation axis, the classification label may be garbage bag unbroken garbage, pruned branches, or futon (the number of label types may be increased) . As another example, when the second evaluation axis is the amount of unburned materials, the classification label may be large, normal, or small (the types of labels may be increased), and the classification label may be large, normal, or small (the types of labels may be increased), and the The amount of unburned material may be ordered according to the amount of unburned material. As another example, if the amount of incinerated ash is used as one second evaluation axis, the classification label may be large, normal, or small (the types of labels may be increased), and the classification label may be large, normal, or small (the types of labels may be increased), and the The data may be arranged in order of the amount of incinerated ash.

The second teacher data generation unit 12b2 generates a combination of the second imaging data 13b2 and process data obtained from various sensors (not shown) installed in the incineration facility 1 and/or the amount of calculation obtained by calculation from the process data. The second teacher data 13c2 may be generated by associating classification labels in each of two or more second evaluation axes that are artificially assigned by the operator based on empirical rules to the combinations. Here, the process data is data that takes into account the time lag between the imaging time of the first imaging data and the response time of the sensor. The time lag may be determined as appropriate based on the response speed of the sensor, the installation position of the sensor within the facility, experiments, simulations, and the operator's empirical rules.

The second imaging data 13b2 may be moving image data within 60 seconds, or may be moving image data within 30 seconds. According to the findings of the present inventors, the combustion state within the incinerator 6 changes every 5 to 10 seconds. Therefore, by using moving image data within 60 seconds (or within 30 seconds) as one segment as the second imaged data 13b1, the combustion state in the incinerator 6 can be classified, predicted, and estimated with higher accuracy. It becomes possible.

Further, the second imaging data 13b2 may be moving image data of 5 seconds or more, or may be 7 seconds or more. While the combustion state in the incinerator 6 changes every 5 to 10 seconds, by using moving image data of 5 seconds or more (or 7 seconds or more) as the second image data 13b2, for example, the incinerator 6 changes every 5 to 10 seconds. It becomes possible to classify, predict, and estimate the combustion state in the engine 6 with higher accuracy.

The second model construction unit 13c2 generates a second trained model by performing machine learning on the second teacher data 13b2 generated by the second teacher data generation unit 12b2 using the second algorithm 13a2. The second algorithm 13a2 used for machine learning includes statistical estimation using maximum likelihood classification, Boltzmann machine, neural network (NN), support vector machine (SVM), Bayesian network, sparse regression, decision tree, and random forest. It may include at least one of boosting, reinforcement learning, and deep learning.

The second image analysis unit 13d2 inputs the new second image data (visible light image data) inside the incinerator 6 acquired by the second image data acquisition unit 13a2, and generates the image data generated by the second model construction unit 13c2. Using the second trained model, evaluation results on each of two or more second evaluation axes (for example, flame condition and amount of combustible material) are obtained as output data. The second image analysis unit 13d2 normalizes (scores) the evaluation results on each of the two or more second evaluation axes into a numerical range of 0 to 100, for example, and converts the numerical values (scores) within the range into output data. You may also obtain it as

The combustion state determination unit 12e determines the current combustion state in the incinerator 6 based on the evaluation results (scores) on each of the two or more second evaluation axes by the second image analysis unit 13d2. The combustion status determination result may be a label indicating the characteristics of the combustion status (over-combustion, thick garbage layer combustion, thin garbage layer combustion, garbage withered garbage, low-quality garbage combustion, etc.), or a numerically modified version of the label. It may be.

The combustion state determination unit 12e places the set of evaluation results on each of the two or more second evaluation axes onto a predetermined combustion state determination map 13d whose coordinate axes are the two or more second evaluation axes. The current combustion state within the incinerator 6 may be determined by mapping.

FIG. 4 is a diagram showing an example of the combustion state determination map 13d. In the combustion state determination map 13d shown in FIG. It is divided into six areas: thin combustion zone, good combustion (normal) zone, thick garbage combustion zone, and low-quality garbage combustion zone. The combustion state determination unit 12e uses the set (X, Y) of the evaluation results on the second evaluation axes when the amount of combustible material and the flame state are two different second evaluation axes to determine the combustion state. It is mapped on the map 13d. For example, as shown in FIG. 4, a set (X, Y ) is mapped within the good combustion (normal) zone, the combustion state determination unit 12e determines that the current combustion state is "good combustion (normal)."

The combustion state determination unit 12e corrects the combustion state determination result according to process data obtained from various sensors (not shown) installed in the facility and/or the amount of calculation obtained by calculation from the process data. Good too. For example, if the combustion state determination result is "low-quality waste combustion" and the furnace outlet temperature or evaporation amount difference value PV-SV satisfies a predetermined condition, the combustion state determination result is It may be corrected to "good combustion (normal)".

Depending on the combustion state determination result, the combustion state determination unit 12e may display an alert such as that it is difficult to maintain the performance management index value within a predetermined range on a display (not shown), or may display an alert using audio or vibration. , the alarm may be issued using light, etc.

Next, an example of an information processing method by the information processing apparatus 10 having such a configuration will be described. FIG. 10 is a flowchart illustrating an example of an information processing method.

As shown in FIG. 10, first, before the incineration facility 1 is operated, the second teacher data generation unit 12b2 generates past visible light imaging data (second imaging data) inside the incinerator 6 captured by the visible light imaging device 72. By linking the classification labels of two or more second evaluation axes that are factors for determining the combustion state, which are artificially assigned by operators based on empirical rules, to the data 13b2), the second training data 13c2 is generated (step S31).

Next, the second model construction unit 13c2 generates a second learned model by performing machine learning on the second teacher data 13b2 generated by the second teacher data generation unit 12b2 using the second algorithm 13a2. (Step S32).

Next, while the incineration facility 1 is operating, the second imaged data acquisition unit 12a2 acquires new visible light imaged data inside the incinerator 6 captured by the visible light imaging device 72 as second imaged data 13b2 (step S33).

Next, the second image analysis unit 13d2 inputs the new second image data (visible light image data) inside the incinerator 6 acquired by the second image data acquisition unit 13a2, and generates the image data by the second model construction unit 13c2. Using the second learned model, evaluation results on each of two or more second evaluation axes (for example, amount of combustible material and flame state) are obtained as output data (step S34).

Then, the combustion state determination section 12e determines the current combustion state in the incinerator 6 based on the evaluation result on the second evaluation axis by the second image analysis section 13d2 (step 35).

In step S35, the combustion state determination unit 12e converts a set of evaluation results on each of the two or more second evaluation axes into a predetermined combustion state determination map having the two or more second evaluation axes as coordinate axes. By mapping onto 13d, the current combustion state within the incinerator 6 may be determined (see FIG. 4).

Thereafter, the instruction section 12f transmits an operation instruction to the crane control device 50 and/or the combustion control device 20 based on the determination result of the combustion state determination section 12e (step S16).

According to the present embodiment as described above, by determining the combustion state based on the evaluation results on each of the two or more second evaluation axes, the combustion state is determined based on the evaluation result on one evaluation axis. The combustion state inside the incinerator 6 can be classified, predicted, and estimated with higher accuracy than when determining the combustion state. Furthermore, by determining the combustion state by mapping the evaluation results for each of the two or more second evaluation axes onto the combustion state determination map 13d, the evaluation results for each of the two or more second evaluation axes can be determined. It is possible to intuitively grasp the relationship between the combustion state and the combustion state that is the determination result, and it is also possible to quickly determine the combustion state.

Note that in the embodiment described above, a part of the processing of the control unit 12 may be performed not on the information processing apparatus 10 but on a cloud server different from the information processing apparatus 10. A part of the storage unit 13 may be located not on the information processing device 10 but on a cloud server separate from the information processing device 10.

For example, the processing of the first teacher data generation unit 12d1 and/or the second teacher data generation unit 12d2 may be executed on a cloud server, and the first teacher data 13c1 and/or the second teacher data 13c2 may be generated. Processing by the first model construction unit 12c1 and/or the second model construction unit 12c2 may be executed on a cloud server to construct a trained model. Further, the processing of the first image analysis unit 12d1 and/or the second image analysis unit 12d2 may be executed on the cloud server using a learned model constructed on the cloud server, or the learned model ( The information processing device 10 may download the learned parameters) from the cloud server, and use this within the information processing device 10 to execute the processing of the first image analysis unit 12d1 and/or the second image analysis unit 12d2. .

Although the embodiments and modifications of the present invention have been described above by way of example, the scope of the present invention is not limited thereto, and may be modified or transformed according to the purpose within the scope of the claims. is possible. Moreover, each embodiment and modification example can be combined as appropriate within a range that does not conflict with the processing contents.

Furthermore, although the information processing device 10 according to the embodiment of the present invention may be configured by one or more computers, it is also possible to use a program for realizing the information processing device 10 on one or more computers and a non-temporary implementation of the program. Computer-readable recording media that have been recorded are also subject to protection in this case.

1 Incineration facility 2 Boiler 3 Garbage pit 4 Hopper 5 Crane 6 Incinerator 71 Infrared imaging device 72 Visible light imaging device 10 Information processing device 11 Communication section 12 Control section 12a1 First imaging data acquisition section 12a2 Second imaging data acquisition section 12b1 1 teacher data generation section 12b2 second teacher data generation section 12c1 first model construction section 12c2 second model construction section 12d1 first image analysis section 12d2 second image analysis section 12e combustion state determination section 12f instruction section 12g process data acquisition section 13 Storage unit 13a1 First algorithm 13a2 Second algorithm 13b1 First image data 13b2 Second image data 13c1 First teacher data 13c2 Second teacher data 13d Combustion state determination map 20 Combustion control device 21 Platform 22 Transport vehicle 50 Crane control device

Claims (25)

First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a first image analysis unit that performs evaluation on the first evaluation axis by inputting new first imaging data in the incinerator using a first learned model that has been machine-learned;
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. a second image analysis unit that performs evaluation on the second evaluation axis by inputting new second imaging data in the incinerator using a second learned model that has been machine-learned;
By mapping the evaluation results on each of the first evaluation axis and the second evaluation axis onto a predetermined combustion state determination map whose coordinate axes are the first evaluation axis and the second evaluation axis, the current combustion state can be determined. a combustion state determination unit that determines the state;
An information processing device comprising: The first imaging device is an infrared imaging device, and the second imaging device is a visible light imaging device.
The information processing device according to claim 1, characterized in that: The first evaluation axis includes at least one of the following: quantity of materials to be burned, quality of materials to be burned, type of materials to be burned, temperature of materials to be burned, amount of combustion gas generated, and temperature of a wall of an incinerator. ,
The information processing device according to claim 2, characterized in that: The second evaluation axis includes the flame condition, the position of the combustion completion point, the shape of the combustion completion point, the amount of combustible material, the quality of combustible material, the type of combustible material, the amount of unburned material, and the amount of incinerated ash. including at least one of
The information processing device according to claim 2 or 3, characterized in that: The first imaging data is moving image data within 60 seconds, and/or
The second imaging data is moving image data within 60 seconds.
The information processing device according to any one of claims 1 to 4, characterized in that: The red first imaging data is moving image data of 5 seconds or more, and/or
The red second imaging data is video data of 5 seconds or more,
The information processing device according to any one of claims 1 to 5. The classification label in the first training data is at least one of a label indicating which of a plurality of predetermined classification items it falls under, and a relative order among the plurality of first imaging data. , and/or
The classification label in the second training data is at least one of a label indicating which of a plurality of predetermined classification items it falls under, and a relative order among the plurality of second imaging data. ,
The information processing device according to any one of claims 1 to 6. The first image analysis unit performs a combination of first imaging data in the incinerator and process data obtained from a sensor installed in the facility and/or a calculation amount obtained by calculation from the process data. New first imaging data in the incinerator and new process data are generated using a first learned model obtained by machine learning the first teaching data generated by assigning a classification label in the first evaluation axis. and/or performs evaluation on the first evaluation axis by inputting a combination with the amount of calculation obtained by calculation from the process data, and/or
The second image analysis unit performs a combination of second imaging data in the incinerator and process data obtained from a sensor installed in the facility and/or a calculation amount obtained by calculation from the process data. New second imaging data in the incinerator and new process data are generated using a second learned model obtained by machine learning the second training data generated by assigning classification labels in the second evaluation axis. and/or performing an evaluation on the second evaluation axis using a combination of the amount of calculation obtained by calculation from the process data as input;
The information processing device according to any one of claims 1 to 7, characterized in that: The process data is
The data takes into account a time lag between the imaging time of the first imaging data and/or the second imaging data and the response time of the sensor,
9. The time lag is determined based on at least one of a response speed of the sensor, an installation position of the sensor within the facility, an experiment, a simulation, and an operator's rule of thumb. Information processing device. The first imaging data inside the incinerator captured by the first imaging device is given a classification label in each of two or more first evaluation axes, which are elements for determining the combustion state. a first image analysis unit that performs evaluation on each of the two or more first evaluation axes by inputting new imaging data in the incinerator using a first learned model obtained by machine learning one teacher data;
The current combustion state can be determined by mapping the evaluation results on each of the two or more first evaluation axes onto a predetermined combustion state determination map that uses the two or more first evaluation axes as coordinate axes. a combustion state determination unit that determines;
An information processing device comprising: the first imaging device is an infrared imaging device;
The information processing device according to claim 10. The first evaluation axis includes at least one of the following: quantity of materials to be burned, quality of materials to be burned, type of materials to be burned, temperature of materials to be burned, amount of combustion gas generated, and temperature of a wall of an incinerator. ,
The information processing device according to claim 11. the first imaging device is a visible light imaging device;
The information processing device according to claim 10. The first evaluation axis includes the flame condition, the position of the combustion completion point, the shape of the combustion completion point, the amount of combustible material, the quality of the combustible material, the type of combustible material, the amount of unburned material, and the amount of incinerated ash. including at least one of
14. The information processing device according to claim 13. The combustion state determination unit corrects the combustion state determination result according to process data obtained from a sensor installed in the facility and/or a calculation amount obtained by calculation from the process data.
The information processing device according to any one of claims 1 to 14. The combustion state determination unit displays or issues an alert depending on the combustion state determination result.
The information processing device according to any one of claims 1 to 15. The information according to any one of claims 1 to 16, further comprising an instruction section that transmits an operation instruction to a crane control device and/or a combustion control device based on the determination result of the combustion state determination section. Processing equipment. The algorithms used for the machine learning include maximum likelihood classification, Boltzmann machine, neural network (NN), support vector machine (SVM), Bayesian network, sparse regression, decision tree, statistical estimation using random forest, and boosting. , reinforcement learning, and deep learning.
The information processing device according to any one of claims 1 to 17. First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a first model construction unit that generates the first learned model by machine learning; and/or
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. The information processing apparatus according to any one of claims 1 to 9, further comprising a second model construction unit that generates the second learned model by performing machine learning. The first imaging data inside the incinerator captured by the first imaging device is given a classification label in each of two or more first evaluation axes, which are elements for determining the combustion state. 15. The information processing apparatus according to claim 10, further comprising a first model construction unit that generates the first trained model by performing machine learning on one teacher data. The information processing device according to claim 17;
The crane control device controls a crane that stirs or transports waste based on an operation instruction transmitted from the information processing device, and/or the combustion control device controls combustion of waste in an incinerator. and,
A system characterized by: An information processing method performed by a computer, the method comprising:
First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a step of performing an evaluation on the first evaluation axis by inputting new first imaging data in the incinerator using a first learned model that has been machine-learned;
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. a step of performing an evaluation on the second evaluation axis by inputting new second imaging data in the incinerator using a second learned model that has been machine-learned;
By mapping the evaluation results on each of the first evaluation axis and the second evaluation axis onto a predetermined combustion state determination map whose coordinate axes are the first evaluation axis and the second evaluation axis, the current combustion state can be determined. a step of determining the state;
An information processing method characterized by comprising: An information processing method performed by a computer, the method comprising:
The first imaging data inside the incinerator captured by the first imaging device is given a classification label in each of two or more first evaluation axes, which are elements for determining the combustion state. a step of performing evaluation on each of the two or more first evaluation axes by inputting new first imaging data in the incinerator using a first learned model obtained by machine learning one teacher data;
The current combustion state can be determined by mapping the evaluation results on each of the two or more first evaluation axes onto a predetermined combustion state determination map that uses the two or more first evaluation axes as coordinate axes. a step of determining;
An information processing method characterized by comprising: to the computer,
First training data generated by assigning a classification label in at least one first evaluation axis, which is an element for determining the combustion state, to the first imaged data inside the incinerator captured by the first imaging device. a step of performing an evaluation on the first evaluation axis by inputting new first imaging data in the incinerator using a first learned model obtained by machine learning;
Second training data generated by assigning a classification label in at least one second evaluation axis, which is an element for determining the combustion state, to the second imaged data inside the incinerator captured by the second imaging device. performing an evaluation on the second evaluation axis by inputting new second imaging data in the incinerator using a second learned model that has been machine-learned;
By mapping the evaluation results on each of the first evaluation axis and the second evaluation axis onto a predetermined combustion state determination map whose coordinate axes are the first evaluation axis and the second evaluation axis, the current combustion state can be determined. a step of determining the state;
An information processing program that executes. to the computer,
The first imaging data inside the incinerator captured by the first imaging device is given a classification label in each of two or more first evaluation axes, which are elements for determining the combustion state. a step of performing evaluation on each of the two or more first evaluation axes by inputting new first imaging data in the incinerator using a first learned model obtained by machine learning one teacher data;
The current combustion state can be determined by mapping the evaluation results on each of the two or more first evaluation axes onto a predetermined combustion state determination map that uses the two or more first evaluation axes as coordinate axes. a step of determining;
An information processing program that executes.