Dehydration state determination device, dehydration state determination method, and dehydration control method for wet dust

Abstract

The present invention provides a means for easily determining the dewatering status of dewatered cakes while reducing the workload on factory workers during the dewatering process. [Solution] A dehydration state determination device for determining the dehydration state of a dehydrated cake, comprising: an acquisition unit 6Ca that acquires an image of a region of the dehydrated cake that can be imaged; and a dehydration state determination unit 6Cb that determines the dehydration state from the image acquired by the acquisition unit 6Ca, using a dehydration state determination model 12 which is trained using the image of the dehydrated cake as input and outputs information on the dehydration state of the dehydrated cake as input.

Description

Technical Field

[0001] The present invention relates to a technique for determining the dehydration state of a dewatered cake generated by a dehydrator, and dehydration control of wet dust using that technique. The dewatered cake targeted by the present invention is, for example, a dewatered cake obtained by dehydrating wet dust. However, the dewatered cake targeted by the present invention is not limited to the dewatered cake formed by dehydrating wet dust.

Background Art

[0002] In the steel manufacturing process, dust containing metal is recovered from exhaust gas generated in various processes, and the dust is reused. For example, as dust collected from the exhaust gas of a blast furnace, there are powdery dry dust and sludge-like wet dust. The dry dust is recovered by a dry dust collector. Also, the wet dust is recovered by a wet dust collector.

[0003] The dry dust has little moisture and is excellent in handling properties. Therefore, the dry dust is reused as a sintering raw material or the like. On the other hand, the wet dust has a high moisture content. Therefore, when reusing the wet dust, it is necessary to remove a large amount of moisture from the wet dust in advance. Thus, the sludge-like wet dust is dehydrated by a dehydrator to form a dewatered cake. For example, a filter press dehydrator is used as the dehydrator. The filter press dehydrator dehydrates the wet dust by pressurizing and squeezing it, thereby obtaining a dewatered cake from which moisture has been removed.

[0004] In such a dehydrator, the dehydration performance changes depending on the time for performing the dehydration treatment (dehydration time). Also, when performing the dehydration treatment, the dehydration time is usually set to a fixed value determined in advance.

[0005] In this process, setting a longer dewatering time can improve dewatering performance. However, the longer the dewatering time, the more easily the filter wears out. Also, the longer the processing time, the lower the productivity. On the other hand, setting a shorter dewatering time may result in insufficient dewatering. Insufficient dewatering can lead to problems such as decreased productivity in the next process, dust adhering to the equipment, and poor raw material handling.

[0006] Furthermore, the appropriate dewatering time varies depending on the properties of the wet dust. Therefore, if the properties of the wet dust recovered from the blast furnace change due to changes in the blast furnace's operating conditions, it is preferable to set an appropriate dewatering time according to those properties. However, if workers have to visually check the state of the dewatered cake and manually readjust the dewatering time each time the properties of the wet dust change, the dewatering process may become cumbersome.

[0007] To address these challenges, for example, the technologies described in Patent Document 1 and Patent Document 2 have been proposed. Patent Document 1 detects the concentration and amount of sludge supplied to a dewatering machine, as well as the amount of filtrate discharged from the dewatering machine. Patent Document 1 also discloses a technology for automatically setting the dewatering time based on these detected values.

[0008] Furthermore, Patent Document 2 measures the amount injected during injection filtration and the amount of filtrate discharged from the dewatering machine. Patent Document 2 also discloses a technology for controlling the operation of the dewatering machine by comparing the calculated concentrated sludge concentration with a set value. [Prior art documents] [Patent Documents]

[0009] [Patent Document 1] Japanese Patent Application Publication No. 54-7673 [Patent Document 2] Japanese Patent Application Publication No. 5-49818 [Overview of the project] [Problems that the invention aims to solve]

[0010] In the technologies described in Patent Documents 1 and 2, the moisture content of the dewatered cake is indirectly estimated from measured values ​​such as the amount of filtrate discharged from the dewatering machine, and the dewatering time is controlled accordingly. In other words, the technologies described in Patent Documents 1 and 2 do not directly determine the dewatered state of the cake. Therefore, depending on the changing properties of the wet dust, dewatering may be insufficient.

[0011] Here, as a means of directly measuring the moisture content of dehydrated cake, the application of a near-infrared moisture meter, which is used for measuring the moisture content of powdered raw materials, can be considered. However, with a near-infrared moisture meter, it is necessary to determine a calibration curve in advance for each component of the raw material being measured.

[0012] However, in the case of wet dust, its composition changes depending on the operating conditions of the blast furnace, making it difficult to ascertain the composition of the wet dust in real time. Furthermore, it is necessary to maintain a constant distance between the moisture meter and the surface of the raw material being measured. However, the surface of the dewatered cake discharged from the dewatered cake is not flat. Therefore, it is difficult to apply a moisture meter to measure the dewatering state of the dewatered cake.

[0013] Therefore, currently, factory workers regularly visually inspect the surface condition of the dewatered cake and adjust the operating conditions of the dewatering machine as needed. However, this process is time-consuming.

[0014] The present invention has been made in view of the above-mentioned points. Its purpose is to enable optimal dewatering treatment in accordance with the changing properties of dewatered cake such as wet dust. Furthermore, one of the purposes of the present invention is to provide a means for easily determining the dewatering state of the dewatered cake while reducing the burden on workers during the optimal dewatering treatment. [Means for solving the problem]

[0015] To solve the problem, one aspect of the present invention is a dehydration state determination device for determining the dehydration state of a dewatered cake, comprising an image acquisition unit that acquires an image of an area where the dewatered cake can be imaged, and a dehydration state determination model that is a model trained using the image of the dewatered cake, which inputs the image and outputs information on the dehydration state of the dewatered cake. The dehydration state determination device has a dehydration state determination unit that determines the dehydration state from the image acquired by the image acquisition unit.

Advantages of the Invention

[0016] According to an aspect of the present invention, it becomes possible to easily determine the dehydration state from an image of a dewatered cake imaged by at least one camera without the need to additionally install a plurality of types of sensors. As a result, according to an aspect of the present invention, for example, even when the properties of the dewatered cake such as wet dust change, the dehydrator can be optimally operated while reducing the burden on factory workers to visually check the state of the dewatered cake regularly.

Brief Description of the Drawings

[0017] [Figure 1] It is a schematic diagram for explaining the configuration of a dehydration facility (dewatered cake manufacturing facility) according to an embodiment based on the present invention. [Figure 2] It is a diagram for explaining the configuration of a dehydration state determination device according to an embodiment based on the present invention. [Figure 3] It is a diagram for explaining an example of learning data for model generation.

Embodiments for Carrying Out the Invention

[0018] Next, embodiments of the present invention will be described with reference to the drawings. In the following description, as the dewatered cake to be determined, a dewatered cake formed by dehydrating wet dust will be described as an example. However, the dewatered cake to which the present invention is applicable is not limited to the dewatered cake generated from wet dust.

[0019] ​FIG. 1 is a diagram for explaining the dehydration equipment of the present embodiment. The dehydration equipment is equipment for manufacturing a dehydration cake. As shown in FIG. 1, this dehydration equipment includes a wet dust collector 1, a thickener 2, a dehydrator 3, a belt conveyor 4, a camera 5, a dehydration state determination device 6, and a control device 7 for the dehydrator 3.

[0020] <Wet dust collector 1> The wet dust collector 1 is continuously supplied with exhaust gas from a blast furnace and collects wet dust in the exhaust gas. The wet dust collector 1 sequentially supplies the collected wet dust to the thickener 2.

[0021] <Thickener 2> The thickener 2 performs a precipitation treatment on the supplied wet dust and supplies the sediment-concentrated wet dust to the dehydrator 3. The wet dust supplied to the dehydrator 3 contains a large amount of moisture and is in a sludge state.

[0022] <Dehydrator 3> The dehydrator 3 of the present embodiment is a device that dehydrates wet dust to generate a dehydration cake. The dehydrator 3 used in the present embodiment is not limited to a specific type of machine, and a known dehydrator may be used. In the present embodiment, it is assumed that a filter press dehydrator 3 is used as the dehydrator 3. The filter press dehydrator 3 pressurizes and squeezes the supplied wet dust to dehydrate it, thereby obtaining a dehydration cake 10 from which moisture has been removed. The removal of moisture by the dehydrator 3 has a positive correlation with the dehydration time. In the present embodiment, the dehydration time of the dehydrator 3 is adjusted according to a command from the control device 7.

[0023] The dehydrator 3 intermittently discharges the generated dehydration cake 10. In this example, the dehydration cake 10 is discharged downward. Note that the dehydration cake 10 may be discharged continuously.

[0024] <Belt conveyor 4> The belt conveyor 4 is an example of a conveyor that conveys the dehydration cake 10. The dewatered cake 10 discharged from the dewatering machine 3 is sequentially placed on the upward belt 4a of the belt conveyor 4 and transported by the belt conveyor 4 to the next process (next step), such as a granulator. The upward belt 4a constitutes a transport path for transporting the dewatered cake 10.

[0025] <Camera 5> Camera 5 is an example of an imaging device. Camera 5 images the dewatered cake 10 being transported by the belt conveyor 4. In this embodiment, as shown in Figure 1, the camera 5 is installed with its imaging axis (imaging unit) facing the upward belt 4a that transports the dewatered cake 10. That is, the camera 5 takes a certain area on the upward belt 4a, which is estimated to be the area through which the dewatered cake 10 passes, as its imaging area (imaging field of view). The camera 5 then images this imaging area.

[0026] The imaging area is set to include, for example, the entire width of the upward belt 4a. The length of the imaging area along the longitudinal direction of the upward belt 4a is set to a length that allows for the capture of a sufficient amount of the dewatered cake 10 to determine its dewatering state. For example, if the dewatered cakes 10 are conveyed intermittently, the imaging area may be set to be larger than the spacing between adjacent dewatered cakes 10, or to a length that is no more than twice the maximum expected length of the dewatered cake 10 along the longitudinal direction of the upward belt 4a, and at least 60% of that maximum length. However, the imaging area is not limited to such a range. The imaging area is not particularly limited as long as it is an area in the captured image where the dewatering state of the dewatered cake 10 can be determined.

[0027] In this embodiment, since the dewatered cake 10 is conveyed intermittently, there is a possibility that all or part of the dewatered cake 10 may not be captured in the image taken by the camera 5. Also, even when the dewatered cake 10 is conveyed continuously, the amount of dewatered cake 10 in the image may be small. In contrast, in this embodiment, if the amount (area) or state of the dewatered cake 10 in the captured image makes it difficult to determine the dewatered state, it is processed as an undeterminable level (a level where there is no dewatered cake 10).

[0028] Furthermore, the cameras 5 may be configured in such a way that multiple units are arranged along the conveying direction of the upward belt 4a to image the areas through which the dewatered cake 10 passes at multiple locations. Alternatively, a floodlight may be installed to illuminate the imaging area on the upward belt 4a. By illuminating the dewatered cake 10 with the floodlight, the camera 5 can more clearly image the condition of the dewatered cake 10, even in a dark indoor environment.

[0029] Camera 5, for example, takes images in response to commands from the dehydration status determination device 6 and supplies the captured images to the dehydration status determination device 6. Camera 5 takes images at predetermined intervals, for example. Commands from the dehydration status determination device 6 include, for example, start commands and stop commands for camera 5, and imaging command commands that instruct camera 5 to take images. In this example, camera 5 is configured to take images continuously at predetermined time intervals when it receives a start command, and to stop taking images when it receives a stop command. Camera 5 may also be a video camera (moving image capture device). Alternatively, the camera 5 may store the images it captures in a database or other storage device, and the dehydration state determination device 6 may refer to that storage device to obtain images of the dehydrated cake.

[0030] Camera 5 is connected to a network, for example. Images captured by camera 5 are transmitted to the dewatering state determination device 6 via the network. The network may be, for example, a LAN (Local Area Network) or the Internet. In this embodiment, not only camera 5 and the dewatering state determination device 6, but also various control devices 7 are connected to the network, enabling the transmission and reception of information between these devices through communication. The dewatering state determination system may also consist of the dewatering state determination device 6 and camera 5. Furthermore, the dewatering state determination system may be configured with other devices (for example, a higher-level control device).

[0031] <Model generation unit 11> The model generation unit 11 is a processing unit that generates a dehydration state determination model 12 used in the determination of the dehydration state determination unit. The model generation unit 11 may be configured as part of the dehydration state determination device 6. The dehydration state determination model 12 generated by the model generation unit 11 is stored in advance in a storage device, such as the storage unit 6B of the dehydration state determination device 6. The model generation unit 11 performs a process to generate a dewatering state determination model 12 before the dewatering state determination device 6 performs a process to determine the dewatering state of the dewatered cake 10.

[0032] In this embodiment, a sufficient number of training data sets are prepared in advance, each set consisting of a training image, which is an image of the dewatered cake 10 being transported by the upward belt 4a, and a level (label) that represents the state of the dewatered cake in the training image. That is, each training data set consists of a "training image and level" pair.

[0033] Figure 3 is an example of training data. Prior to the experiment, for example, a worker visually inspects the state of the dehydrated cake in each training image to determine the level at which to represent the state of the dehydrated cake. For example, a label (0-3 in the example in Figure 3) corresponding to the absence or level of dehydration of the dehydrated cake 10 was determined and set. In this example, the dehydration state of the dehydrated cake 10 was identified by the degree of light reflection (shininess) of the dehydrated cake 10 in the image.

[0034] Next, let me explain the levels. In this embodiment, images in which the dewatered cake 10 is not visible or is barely visible are labeled with "0" corresponding to "No dewatered cake 10". Furthermore, images with weak gloss and good dehydration were labeled "1" corresponding to "Good". Images with gloss and where the dehydrated cake 10 is cohesive, and where the dehydration is within an acceptable range, were labeled "2" corresponding to "Acceptable". Images with strong gloss and where the dehydrated cake 10 is spread out, and where there is excessive moisture, were labeled "3" corresponding to "Unacceptable". As described above, in this example, the level was set so that the higher the moisture content, the higher the value. In addition, a unique numerical value was assigned as the level for images that could not be used for judgment.

[0035] In this example, 2421 images labeled 0, 177 images labeled 1, 365 images labeled 2, and 236 images labeled 3 were prepared as training data, and machine learning was performed. A known machine learning method can be applied. The generated dehydration status determination model 12 is a model that takes image data to be determined as input and outputs information about the dehydration status. In this example, the dehydration status information will be output as a level value (one of 0 to 3).

[0036] In this example, validation data was also prepared separately to evaluate the generated dehydration state determination model 12. This validation data consisted of 1211 images labeled 0, 89 images labeled 1, 183 images labeled 2, and 118 images labeled 3. Each image from the validation data was input into the generated dehydration state determination model 12, and the output level values ​​were verified to confirm that the model could make accurate determinations.

[0037] <Dehydration State Determination Device 6> The dewatering state determination device 6 is a device for determining the dewatering state of the dewatered cake 10 based on the image captured by the camera 5 (imaging device). The dewatering state determination device 6 is configured, for example, as part or all of a computer. That computer may be a server computer, or a portable computer such as a laptop or tablet.

[0038] As shown in Figure 2, the dehydration state determination device 6 of this embodiment comprises a communication unit 6A, a storage unit 6B, and a control unit 6C. The control unit 6C also comprises an acquisition unit 6Ca, a dehydration state determination unit 6Cb, and an output unit 6Cc. The dehydration state determination device 6 also includes a CPU and other components as a computer processing unit. Furthermore, the storage unit 6B and the control unit 6C may be configured on a single hardware device or on separate hardware devices.

[0039] [Communications Section 6A] The communication unit 6A comprises one or more communication modules connected to a network. The communication unit 6A may include communication modules that support mobile communication standards such as 4G (4th Generation) and 5G (5th Generation). The communication unit 6A may also include communication modules that support wired or wireless LAN standards. In this embodiment, the communication unit 6A is configured to exchange information with the camera 5 via communication.

[0040] [Storage section 6B] The storage unit 6B has one or more memories. Examples of memories include semiconductor memory, magnetic memory, or optical memory. The storage unit 6B is not limited to these and can be composed of any memory. In this example, the storage unit 6B is built into the dehydration state determination device 6, but the storage unit 6B may be configured to be accessed externally by the dehydration state determination device 6 via any interface.

[0041] The storage unit 6B stores various data used in various calculations performed by the control unit 6C. The storage unit 6B may also store the results and intermediate data of the various calculations performed by the control unit 6C. The storage unit 6B may include a database, as described later. Furthermore, the control unit 6C may be composed of a program, and that program may be stored in the storage unit 6B.

[0042] [Control Unit 6C] The control unit 6C consists of one or more processors. These processors may be, for example, general-purpose processors or dedicated processors specialized for specific processing. However, they are not limited to these. The control unit 6C can be composed of any processor. The control unit 6C is a system that controls the overall operation of the dehydration state determination device 6. The control unit 6C may be composed entirely or partially of software. In this example, one or more programs used to control the operation of the dehydration state determination device 6 are stored in the storage unit 6B. When the programs stored in the storage unit 6B are read by the processor of the control unit 6C, the control unit 6C is made to function as an acquisition unit 6Ca, a dehydration state determination unit 6Cb, and an output unit 6Cc.

[0043] [Acquisition part 6Ca] The acquisition unit 6Ca acquires images from the camera 5 that has captured an area of ​​the dehydrated cake 10 that can be imaged. The acquisition unit 6Ca constitutes the image acquisition unit 6Ca. The camera 5 may be included as part of the image acquisition unit 6Ca.

[0044] The acquisition unit 6Ca in this embodiment acquires images of the dewatered cake 10 captured by the camera 5. The acquisition unit 6Ca in this embodiment receives images continuously from the camera 5 by transmitting a start command to the camera 5. The acquisition unit 6Ca supplies images from the received images at predetermined intervals to the dewatering state determination unit 6Cb. The acquisition unit 6Ca may be configured to transmit an imaging command to the camera 5 and acquire an image from the camera 5 each time. For example, it may be set to perform imaging by the camera 5 in synchronization with the discharge of the dewatered cake 10 from the dewatering machine 3. The images captured by camera 5 may be sequentially stored in a database accessible by the dehydration state determination device 6, for example.

[0045] The image acquisition interval from camera 5 is set to a constant interval. The imaging interval should be adjusted according to the imaging area (imaging field of view) and the conveying speed of the belt conveyor 4 so that all of the dewatered cakes 10 being transported are captured without fail. For example, consider a case where the conveying speed of the belt conveyor 4 is 1 m / s and imaging is performed with an imaging field of view of 2 m in the longitudinal direction of the belt. In this case, images captured at intervals of 2s are used. This increases the likelihood that the dewatered cakes 10 will be visible in the image, and the dewatering state can be determined more accurately from the image.

[0046] [Dehydration status determination unit 6Cb] The dehydration state determination unit 6Cb uses the dehydration state determination model 12 to obtain information for determining the dehydration state from the image acquired by the image acquisition unit 6Ca. In this example, the information is expressed as a level value.

[0047] The dehydration state determination unit 6Cb determines the dehydration state of the dehydrated cake 10 based on the acquired information (level value). However, if the level value is 0, it is ignored. Here, the dewatered cake 10 is generally discharged onto the conveyor belt 4 intermittently and transported. Therefore, the presence or absence of the dewatered cake 10 in the image may change frequently. Consequently, if the result of determining the dewatering level of images captured at regular intervals is output as is, the "dewatering level" for cases where there is no dewatered cake 10 and the "dewatering level" for cases where there is a dewatered cake 10 will be output sequentially, making it difficult to grasp the dewatering level. Therefore, in the case of a level of "no dewatered cake 10" (=0), that level is ignored. Then, if the level is not 0, the dehydration state determination unit 6Cb uses that level value as information for determining the dehydration state. Furthermore, even with dehydrated cakes processed for the same dehydration time, there may be variations in the levels obtained consecutively. If this data is then output directly to the control device 7, the control value (dehydration time) may change frequently. To prevent this, it is preferable to use the average value of multiple levels obtained consecutively as information for determining the dehydration state. For example, the average value can be calculated from the most recent history of multiple measurements. In addition, other statistical methods such as the median may be used.

[0048] [Output section 6Cc] The output unit 6Cc outputs the determination result, including the dewatering status of the dewatered cake 10, as determined by the dewatering status determination unit 6Cb, to another device. The output unit 6Cc may also be configured as part of the dewatering status determination unit 6Cb. In the example shown in Figure 2, the output unit 6Cc outputs the determination result (dewatering status information) to the control device 7. (Control device 7 of the dehydrator 3) The control device 7 controls the operation of the dewatering machine 3, etc., based on the dewatering status information determined by the dewatering status determination device 6, so that a dewatered cake 10 in the target dewatering state can be obtained. The control device 7 in this embodiment controls the operation of the dewatering machine 3 based on information about the dewatering level of the dewatered cake 10. For example, if the level is "1", the control ends without changing the control information. If the level is "2" or "3", the dewatering time is changed to a value longer than the current set time, according to the level value. Furthermore, a value representing an excessive dehydration state of the dehydrated cake 10 may be set as the level of dehydration, and if this value is reached, the dehydration time may be changed to a value shorter than the current set time. In this case, the dewatering time may be controlled by an operator who manually sets and changes the dewatering time according to the dewatering status information (level) determined by the dewatering status determination device 6. (Operation and other functions) As described above, in this embodiment, the presence or absence of the dewatered cake 10 and the level of dewatering, which has been pre-classified, are determined together from the acquired image, so the dewatering state of the dewatered cake 10 can be determined simply and directly. In this embodiment, machine learning-based image classification is used. Therefore, it is possible to determine the dehydration state with higher accuracy compared to using other image processing methods (for example, judgment based on average brightness values). For example, with judgment methods based on average brightness values, it is difficult to determine the presence or absence of dehydrated cake and how the dehydrated cake has spread. In contrast, with machine learning-based image classification, there is no limit to the number of categories to be divided into levels, and the dehydration state can be determined by the desired classification.

[0049] Furthermore, according to this embodiment, since the state of the dewatered cake 10 is directly determined using images from at least one camera 5, the dewatering state can be determined with greater accuracy compared to a configuration that indirectly estimates the state of the dewatered cake 10 from measured values ​​such as the amount of filtrate discharged from the dewatering machine 3.

[0050] Furthermore, in this embodiment, the dewatering state is determined from each image of the transported dewatered cake 10 captured at predetermined intervals by the camera 5. In this case, there is a possibility that the dewatered cake 10 may not be captured in the image. However, since the system also determines the state in which the dewatered cake 10 is absent as part of the determination information, even with a simple configuration, it is possible to accurately determine the level of dewatering in the image using the dewatering state determination model 12 (trained model).

[0051] (modified version) While embodiments of this disclosure have been described based on the drawings and examples, it should be noted that those skilled in the art will find it easy to make various modifications or alterations based on this disclosure. Therefore, it should be noted that these modifications or alterations are included within the scope of this disclosure. For example, the functions included in each component or process can be rearranged in a logically consistent manner, and multiple components or processes can be combined into one or divided. Embodiments relating to this disclosure can also be realized as programs executed by a processor in the device or as storage media recording such programs. These should also be understood to be included within the scope of this disclosure.

[0052] Furthermore, the dehydration state determination device 6 is not a single device, but may consist of multiple devices located in multiple locations that can send and receive data from each other via a network. In other words, it is sufficient for multiple devices connected by a network to function as the dehydration state determination device 6 shown in Figure 2. Therefore, for example, the dehydration state determination device 6 may consist of a single computer in terms of hardware configuration, or it may consist of multiple computers connected by a network. When it consists of multiple computers, the storage unit 6B may be a shared memory accessible by each computer.

[0053] Furthermore, in the above embodiment, the level of dehydration was used as an example of information about the dehydration state, but it is not limited to this. The information about the dehydration state may be any information other than the level of dehydration state, as long as it can be used to determine the dehydration state. For example, the information about the dehydration state may be the moisture content of the dehydrated cake or information related thereto. It is preferable that the information about the dehydration state be expressed in stages. In addition, the information about the dehydration state may be a symbol other than a numerical value, and the method of expressing the information is not particularly limited as long as it can be used to determine the dehydration state.

[0054] (others) This disclosure may also take the following form: (1) Disclosure 1 is a dehydration state determination device for determining the dehydration state of a dehydrated cake, An image acquisition unit that acquires an image of the area of ​​the dehydrated cake that can be imaged, A dehydration state determination unit determines the dehydration state from the image acquired by the image acquisition unit, using a dehydration state determination model that has been trained using images of a dehydrated cake as input and outputs information about the dehydration state of the dehydrated cake as input. A dehydration state determination device having the following features. (2) Disclosure 2 states that the above dehydration state determination model outputs a dehydration state level as information about the dehydration state, The above dehydration levels are expressed numerically. (3) Disclosure 3 has a value that indicates a state in which there is no dehydrated cake in the image, as the level of the dehydration state described above. (4) Disclosure 4 states that the dewatered cakes discharged from the dewatering machine are transported by a conveyor belt. The images used to generate the above model, and the images acquired by the above image acquisition unit, are images of the dewatered cake on the transport path. (5) Disclosure 5 states that the dehydration state determination unit obtains a level of dehydration for each of the multiple images obtained by the image acquisition unit using the dehydration state determination model, and determines the dehydration state by the average of the dehydration state levels obtained by excluding the result of no dehydration cake from the multiple dehydration state levels. (6) Disclosure 6 is a method for determining the dehydration state of a dehydrated cake, We acquire an image of the area where the dehydrated cake can be imaged. A machine learning model is trained using images of a dehydrated cake, and this dehydrated state determination model is used to determine the dehydrated state of the dehydrated cake from the acquired images. A method for determining the dehydration state of a cot. (7) Disclosure 7 describes how wet dust collected from the exhaust gas of a blast furnace is dewatered in a dewatering machine to obtain a dewatered cake, When transporting the dewatered cake discharged from the above-mentioned dewatering machine to the next process via a conveyor belt, the dewatering state of the dewatered cake is determined using the dewatering state determination device described in any one of claims 1 to 5. Based on the determined dehydration state, the dehydration time of the dehydrator is controlled. [Explanation of Symbols]

[0055] 1 Wet type dust collector 2 Sixner 3 Dehydrator 4 Belt conveyor 4a Upward belt 5 Cameras 6. Dehydration status determination device 6A Communications Department 6B Storage section 6C Control Unit 6Ca Image Acquisition Unit 6Cb Dehydration state determination unit 6Cc output section 7 Control device 10 Dehydrated Cake 11 Model Generation Unit 12. Dehydration status determination model

Claims

1. A dehydration state determination device for determining the dehydration state of a dehydrated cake, An image acquisition unit that acquires an image of the area of ​​the dehydrated cake that can be imaged, A dehydration state determination unit determines the dehydration state from the image acquired by the image acquisition unit, using a dehydration state determination model that has been trained using images of a dehydrated cake as input and outputs information about the dehydration state of the dehydrated cake as input. A dehydration state determination device having the following features.

2. The above dehydration status determination model outputs a dehydration status level as information about the dehydration status. The above dehydration level is expressed numerically. A dehydration state determination device as described in claim 1.

3. The above dehydration level has a value that indicates a state where there is no dehydrated cake in the image. A dehydration state determination device as described in claim 2.

4. The dewatered cakes, after being removed from the dewatering machine, are transported along a conveyor belt. The images used to generate the above model, and the images acquired by the above image acquisition unit, are images of the dewatered cake on the transport path. A dehydration state determination device as described in claim 3.

5. The dehydration state determination unit obtains a dehydration state level for each of the multiple images obtained by the image acquisition unit using the dehydration state determination model, and determines the dehydration state by averaging the dehydration state levels obtained from the multiple dehydration state levels, excluding the result for no dehydrated cake. Dehydration state determination device according to claim 3.

6. A method for determining the dehydration state of a dehydrated cake, We acquire an image of the area where the dehydrated cake can be imaged. A machine learning model is trained using images of a dehydrated cake, and this dehydrated state determination model is used to determine the dehydrated state of the dehydrated cake from the acquired images. A method for determining the dehydration state of a cot.

7. Wet dust collected from blast furnace exhaust gas is dewatered in a dewatering machine to form a dewatered cake. When transporting the dewatered cake discharged from the above-mentioned dewatering machine to the next process via a conveyor belt, the dewatering state of the dewatered cake is determined using the dewatering state determination device described in any one of claims 1 to 5. Based on the determined dehydration state, the dehydration time of the dehydrator is controlled. A method for controlling the dewatering of wet dust.

Images