Operation control system and operation control method for dust removal device based on jellyfish gathering predictions using artificial intelligence

Abstract

To provide an operation control system and an operation control method for a dust removal device based on jellyfish gathering predictions using artificial intelligence, which can efficiently control the operation of the dust removal device even if jellyfish gather around the dust removal device.SOLUTION: An operation control system 10 for a dust removal device based on jellyfish gathering predictions using artificial intelligence of the present invention comprises image acquisition means 20 that is placed on a water intake channel upstream of the dust removal device and takes images of the underwater area, image processing means 40 that converts the images obtained by taking images of the underwater world into image data using an image analysis program and recognizes the jellyfish in the images using artificial intelligence, a water level difference meter 30 that measures the difference in water level before and after the dust removal device, and control means 60 that operates the dust removal device in response to the gathering of the jellyfish when the image processing means detects the jellyfish and when the measurement value of the water level difference meter exceeds a set value.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an operation control system and an operation control method based on jellyfish attack prediction using artificial intelligence for a dust removal device installed in a waterway for taking in industrial water. [Background technology]

[0002] At power plants and other industrial water users, dust removal devices are installed in the waterways that take in cooling water and industrial water. These dust removal devices function to remove foreign objects such as jellyfish and debris that flow into intakes in the sea, rivers, ponds, etc. Power plants and other facilities that require large amounts of water intake use a structure that allows dust removal without clogging by rotating a rotating mesh screen up and down to continuously entangle and capture debris.

[0003] Conventional dust removal devices control operation by measuring the difference in water level before and after the device, and for example, operate at low speed when the water level difference exceeds 200 mm, and operate at high speed when the water level difference exceeds 500 mm to capture and process dust (prior art in Patent Document 1). When seawater is taken in at the location where such dust removal equipment is installed, sudden mass jellyfish outbreaks can occur. If the dust removal equipment's processing capacity is exceeded, the power plant or other facility must reduce the amount of water it takes in, necessitating undesirable measures. Therefore, if jellyfish outbreaks could be predicted in advance, the dust removal equipment could be operated efficiently, without reducing the amount of water taken, and it would be possible to avoid excessive loads on the dust removal equipment due to continuous operation. Conventional camera monitoring only sees the water surface and cannot detect jellyfish below the surface. For this reason, during the summer, when jellyfish attack, workers operate the equipment continuously regardless of whether jellyfish are present, which has led to problems with equipment deterioration over time due to long-term loads. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 5-321228 Summary of the Invention [Problem to be solved by the invention]

[0005] The problem that the present invention aims to solve is, in view of the problems of the above-mentioned conventional technology, to provide an operation control system and an operation control method for a dust removal device based on jellyfish attack prediction using artificial intelligence, which can efficiently control the operation of the dust removal device even if jellyfish attack the dust removal device. [Means for solving the problem]

[0006] As a first means for solving the above problems, the present invention provides an image acquisition means that is arranged on the water intake channel upstream of the dust removal device and takes images of the underwater environment; Image processing means for converting the underwater image obtained by photographing the underwater scene into image data using an image analysis program, and recognizing the jellyfish based on the correlation between the area ratio of the shadow in the image and the subsequent change in the water level difference using artificial intelligence; a water level difference meter that measures the water level difference before and after the dust removal device; The object of the present invention is to provide an operation control system for a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that it is equipped with a control means for operating the dust removal device in response to the jellyfish attack when the jellyfish are detected by the image processing means and when the measurement value of the water level difference meter exceeds a set value. According to the first means, the dust removal device is efficiently controlled to operate at low or high speed to detect the occurrence of jellyfish and follow the infestation, thereby avoiding continuous operation and reducing deterioration of the device over time due to long-term load.

[0007] As a second means for solving the above-mentioned problems, the present invention provides an operation control system for a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that in the first means, the image acquisition means uses sonar that emits sound waves underwater to detect the jellyfish. According to the second means, it is possible to detect jellyfish not only on the surface of the water but also underwater.

[0008] As a third means for solving the above-mentioned problems, the present invention provides an operation control system for a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that, in the first or second means, the system further comprises a learning means for learning the image recognition of the image processing means, the operating status of the dust removal device, and the water level difference of the water level difference meter using artificial intelligence, and changing the set values ​​for operation control. According to the third means, the setting value for starting operation of the dust collector can be set to a value that allows the dust collector to operate efficiently, in other words, to capture large amounts of dust by switching between low and high speed operation without relying on continuous operation.

[0009] As a fourth means for solving the above problems, the present invention provides a method for controlling the operation of a dust removal device based on a jellyfish attack prediction from the detection results by an image analysis program and artificial intelligence installed in a computer using images obtained by photographing underwater, an image acquisition step in which a computer acquires an underwater image; A jellyfish detection step in which a computer converts the image into image data using an image analysis program and detects jellyfish using artificial intelligence based on the correlation between the area ratio of shadows in the image and the subsequent change in water level difference; an increase prediction step in which a computer predicts an increase in jellyfish using artificial intelligence from image data that changes over time; a water level difference acquisition step in which the computer acquires a water level difference before and after the dust removal device; The present invention provides an operational control method for a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that the method includes an operational control step in which a computer operates the dust removal device in response to the jellyfish attack when the jellyfish are detected in the jellyfish detection step and when the measured value in the water level difference acquisition step exceeds a set value. According to the fourth means, the dust removal device is efficiently controlled to operate at low or high speed to detect the occurrence of jellyfish and follow the onslaught, thereby avoiding continuous operation and reducing deterioration of the device over time due to long-term load.

[0010] As a fifth means for solving the above-mentioned problems, the present invention provides a method for controlling the operation of a dust removal device based on jellyfish attack prediction using artificial intelligence, characterized in that in the fourth means, the image acquired in the image acquisition step is an image obtained by detecting the jellyfish by emitting sound waves into the water with a sonar. According to the fifth means, it is possible to detect jellyfish not only on the surface of the water but also underwater.

[0011] As a sixth means for solving the above-mentioned problems, the present invention provides an operation control method for a dust removal device based on jellyfish attack prediction using artificial intelligence, characterized in that in the fourth or fifth means, the computer is further provided with a learning step in which the computer uses artificial intelligence to learn the image recognition in the jellyfish detection step, the operating status of the dust removal device, and the water level difference in the water level difference acquisition step, and changes the setting value in the operation control step. According to the sixth means, the setting value for starting operation of the dust collector can be set to a value that allows the dust collector to operate efficiently, in other words, to capture large amounts of dust by switching between low and high speed operation without relying on continuous operation. [Effects of the Invention]

[0012] According to the present invention, the dust removal device is efficiently controlled to operate at low or high speed to detect the occurrence of jellyfish and follow the infestation, thereby avoiding continuous operation and reducing deterioration of the device over time due to long-term load. [Brief explanation of the drawings]

[0013] [Figure 1] FIG. 1 is a schematic diagram of the configuration of a dust removal device operation control system based on jellyfish attack prediction using the artificial intelligence of the present invention. [Figure 2] FIG. 1 is a block diagram of a dust removal device operation control system based on jellyfish attack prediction using the artificial intelligence of the present invention. [Figure 3] FIG. 1 is a process flow diagram of a dust removal device operation control method based on jellyfish attack prediction using artificial intelligence of the present invention. [Figure 4] 10 is a timing chart of the operation control of the dust remover before and after learning. DETAILED DESCRIPTION OF THE INVENTION

[0014] An embodiment of a dust removal device operation control system and operation control method based on jellyfish attack prediction using artificial intelligence of the present invention will be described in detail below with reference to the drawings.

[0015] [Operation control system for dust removal equipment based on jellyfish attack prediction using artificial intelligence 10] Figure 1 is a schematic diagram of a dust removal device operation control system based on jellyfish attack prediction using artificial intelligence of the present invention. Figure 2 is a block diagram of a dust removal device operation control system based on jellyfish attack prediction using artificial intelligence of the present invention. As shown in the figure, the dust removal device operation control system 10 based on jellyfish attack prediction using artificial intelligence of the present invention includes an image acquisition means 20 located in the water intake channel upstream of the dust removal device and capturing underwater images, an image processing means 40 that converts the captured underwater images into image data using an image analysis program and uses artificial intelligence to recognize jellyfish based on the correlation between the shadow area ratio in the image and the subsequent change in water level difference, a water level difference meter 30 that measures the water level difference before and after the dust removal device, and a control means 60 that operates the dust removal device in response to the jellyfish attack when the image processing means 40 detects a jellyfish or when the measurement value of the water level difference meter exceeds a set value.

[0016] The dust removal device operation control system 10 based on jellyfish attack prediction using artificial intelligence (AI) is composed of a personal computer connected to an image acquisition means 20 and a water level difference meter 30 via wired or wireless communication, and a server equipped with known machine learning software that implements a neural network using a folded neural network algorithm, an AI (artificial intelligence). The personal computer is equipped with a CPU, ROM, RAM, interface, hard disk drive, etc. The CPU controls various processes using the RAM through various programs stored in the ROM. The CPU also functions as a storage means, input means, display means such as a display, and output means such as a printer, thanks to various computer programs stored in the ROM. The server is a computer equipped with a known image analysis program and known AI, and is capable of detecting jellyfish and other debris in image data. It is also configured to learn image recognition, the operating status of the dust removal device, and water level differences.

[0017] The sonar, which serves as the image acquisition means 20, is a device that uses sound waves or ultrasonic waves that propagate through the water to obtain information about jellyfish and other debris in the water. In this embodiment, the sonar is installed upstream of the dust removal device 1. The installation location takes into consideration the need for a predetermined time gap between the detection of a jellyfish and its attack on the dust removal device 1. For example, if the water current is 0.3 m / s and it takes 100 seconds for a jellyfish to attack, the sonar should be installed 30 meters upstream of the dust removal device 1. The sonar is then suspended in seawater from a support member that projects from the ground into the waterway, and images of jellyfish and other debris in the water are captured.

[0018] The water level difference meter 30 is a sensor capable of measuring the water level difference, which is the difference between the water levels upstream and downstream of the dust removal device 1, and in this embodiment, a non-contact radar measurement sensor is used as an example. The non-contact radar measurement sensor emits a radar signal from above the waterway toward the water surface and detects the reflected signal to measure the water level. The water level difference meter 30 can also measure flow velocity, and in this case, it is recommended to install it in a location in the waterway where the flow velocity is high (i.e., in the center of the waterway).

[0019] The image processing means 40 converts the underwater images captured by the image acquisition means 20 into image data using an image analysis program. Jellyfish are then recognized from the shadows in the image data using deep learning, which uses a folding neural network algorithm, an artificial intelligence (AI). In conventional sonar images, jellyfish, fish, etc. appear as shadows. Therefore, whether they affect the dust removal device can only be determined by correlation with the water level difference in the dust removal device 1. Even if the shadow is a fish, it must be identified as a jellyfish and the dust removal device 1 must be operated. Therefore, using AI, if a shadow appears in the image data but the shape of the shadow does not change over time or the water level difference on the screen does not change, it is determined to be a fish or other object that does not affect the water intake, and the data is accumulated and learned. Specifically, jellyfish recognition using AI involves determining the size of the colored area (shadow: jellyfish amount) in the image data to determine the amount of jellyfish. The relationship between the change in the amount of shadow over time and the water level difference in the dust removal device 1 can then be used to identify jellyfish and determine the amount of jellyfish. For example, if the number of jellyfish increases, the area of ​​the shadow in the sonar image will inevitably increase over time, and the amount of jellyfish can be determined by learning from the correlation between this change and the change in the water level difference in the dust removal device.Also, since AI can use other things than jellyfish, such as fish, to move on their own, even if they form a school and appear in the sonar image, the amount of change in the shadow will be different from that of jellyfish, and no change in the water level difference in the dust removal device 1 will be displayed, so it can be determined that they are not jellyfish and can be learned from accumulating data. In this way, instead of humans judging sonar images, AI is used to make mechanical judgments, making it useful for autonomous driving.

[0020] The learning means 50 uses AI to recognize images and learn from accumulated past data on the operating status of the dust remover and the water level difference. The learning means 50 continually accumulates data on the amount of jellyfish determined by image recognition and the subsequent occurrence of water level differences, learning the correlation between changes in the amount of jellyfish and changes in water level difference. Based on the results of this learning, the set water level difference at which low-speed operation of the dust remover 1 begins and the set water level difference at which high-speed operation begins are changed. This allows the artificial intelligence to predict and detect increases and decreases in the number of jellyfish from the results of image recognition by the image processing means 40. This allows the dust remover 1 to be efficiently controlled to switch between low-speed and high-speed operation in response to an attack by jellyfish.

[0021] As an example of learning, target values ​​and initial conditions are set to learn in a direction to reduce operation frequency. The target values ​​are to maintain a water level difference of less than 200 mm and minimize power consumption (operating time). The initial conditions are to operate at low speed if the water level difference is 200 mm or more, and continue until the water level difference is less than 200 mm. If the water level difference is 300 mm or more, to operate at high speed, and continue until the water level difference is less than 200 mm. Operation at low speed occurs when the shadow area in the sonar image is 0% or more, and high speed occurs when the shadow area in the sonar image is 0% or more. Note that the area value fluctuates depending on the learning results. The operation method is set based on correlation learning, assuming a water level difference of less than 200 mm is met, and operation is optimized to reduce the operation frequency of the dust collector 1. The setting for the sonar image area is determined after an introduction test is conducted to actually determine the relationship between the number of jellyfish and the water level difference, etc. After that, learning is performed based on the correlation with the water level difference, and the operation start conditions in the sonar image are changed.

[0022] Examples of the correlation between input and output include inputs such as changes over time in the amount of jellyfish obtained by sonar image analysis, date data, and the water level difference at that time. Outputs include the ON / OFF operation or speed change of the dust removal device 1. Examples of correlation evaluations include data on changes over time in the water level difference from the start to stop of operation, power consumption (operating time), etc. The control means 60 receives an operation command from the learning means 50 and controls the operation of the dust remover 1 to one of low-speed operation, high-speed operation, and stop.

[0023] [Operation control method for dust removal equipment based on jellyfish attack prediction using artificial intelligence] The method of controlling the operation of a dust removal device based on jellyfish attack prediction using artificial intelligence of the present invention detects jellyfish using artificial intelligence according to the processing flow diagram of Figure 3, and controls the operation of the dust removal device based on image recognition and water level difference. (Step 1) In response to commands from a personal computer, sonar captures images of the underwater world continuously or at predetermined intervals, and the images are stored in the personal computer (image acquisition step). The images acquired in the image acquisition step are converted into image data by an image analysis program. Jellyfish recognition using AI then determines that all colored areas (shadows: amount of jellyfish) in the image data are jellyfish, and the amount of jellyfish is determined by recognizing their size. Jellyfish can then be recognized and the amount of jellyfish determined based on the relationship between the change in the amount of shadow over time and the difference in water level in the dust removal device 1 (jellyfish detection step).

[0024] (Step 2) When a jellyfish is detected, the dust remover 1 is operated at a low speed in response to a command from the personal computer (operation control step). (Step 3: Increase prediction step) The artificial intelligence of the learning means 50 predicts or detects an increase in jellyfish from image data that changes over time. (Step 4) When an increase in jellyfish is predicted or detected, the dust remover 1 is operated at high speed in response to a command from the personal computer (operation control step). (Step 5: Water level difference acquisition step) In response to a command from the personal computer, the water level difference meter 30 measures the water levels before and after the dust remover 1, and the water level difference is acquired by the personal computer.

[0025] (Step 6) The learning means 50 determines whether the sonar has detected no jellyfish and whether the water level difference in the dust remover 1 has been restored. If YES, proceed to step 7. If NO, continue with step 4. (Step 7) When the sonar detects no more jellyfish and it is determined that the water level difference in the dust removal device 1 has recovered, the dust removal device 1 is stopped by a command from the personal computer.

[0026] (Step 8: Learning step) The learning means 50 learns image recognition, operating conditions, and water level difference, and changes the set values ​​for the operation control steps, etc. Figure 4 is a timing chart of the dust removal device's operation control before and after learning. The timing to start operation is delayed when the amount of jellyfish generated is high. As a result, the operation time is shortened. This can be said to reflect the results of learning the time difference between jellyfish detection and their arrival at the dust removal device. The amount of jellyfish generated (set value) at which low-speed operation begins is higher before and after learning, and has been changed. In addition, the timing to start high-speed operation is now when the water level difference begins to rise, so the operation time for low-speed operation is longer and the operation time for high-speed operation is shorter. This can be said to reflect the results of learning the change over time in the water level difference in relation to jellyfish detection and generation amount.

[0027] Although the preferred embodiments of the present invention have been described above, the present invention is not limited to the above embodiments and various modifications can be made without departing from the spirit and scope of the present invention. Furthermore, the present invention is not limited to the combinations shown in the embodiments, but can be implemented in various combinations. [Explanation of symbols]

[0028] 1 Dust removal device 10. Dust removal equipment operation control system based on jellyfish attack prediction using artificial intelligence 20 Image acquisition means 30 Water level gauge 40 Image processing means 50 Learning Tools 60 Control Means

Claims

1. an image acquisition means disposed on the water intake channel upstream of the dust removal device to capture images of the underwater environment; Image processing means for converting the underwater image obtained by photographing the underwater scene into image data using an image analysis program, and recognizing the jellyfish based on the correlation between the area ratio of the shadow in the image and the subsequent change in the water level difference using artificial intelligence; a water level difference meter that measures the water level difference before and after the dust removal device; An operation control system for a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that it is equipped with a control means that operates the dust removal device in response to the jellyfish attack when the jellyfish are detected by the image processing means and when the measurement value of the water level difference meter exceeds a set value.

2. An operation control system for a dust removal device based on jellyfish attack prediction utilizing the artificial intelligence according to claim 1, The image acquisition means uses sonar to detect jellyfish by emitting sound waves underwater, and this is an operation control system for a dust removal device based on jellyfish attack predictions using artificial intelligence.

3. An operation control system for a dust removal device based on jellyfish attack prediction utilizing the artificial intelligence according to claim 1 or 2, An operation control system for a dust removal device based on jellyfish attack prediction using artificial intelligence, characterized by comprising a learning means that uses artificial intelligence to learn the image recognition of the image processing means, the operating status of the dust removal device, and the water level difference of the water level difference meter, and changes the setting values ​​for operation control.

4. A method for controlling the operation of a dust removal device based on a jellyfish attack prediction based on the detection results of an image analysis program and artificial intelligence installed in a computer using images obtained by capturing underwater images, an image acquisition step in which a computer acquires an underwater image; A jellyfish detection step in which a computer converts the image into image data using an image analysis program and detects jellyfish using artificial intelligence based on the correlation between the area ratio of shadows in the image and the subsequent change in water level difference; an increase prediction step in which a computer predicts an increase in jellyfish using artificial intelligence from image data that changes over time; a water level difference acquisition step in which the computer acquires a water level difference before and after the dust removal device; A method for controlling the operation of a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that the method includes an operation control step in which a computer operates the dust removal device in response to the jellyfish attack when the jellyfish are detected in the jellyfish detection step and when the measured value in the water level difference acquisition step exceeds a set value.

5. The method for controlling the operation of a dust removal device based on a jellyfish attack prediction using artificial intelligence according to claim 4, A method for controlling the operation of a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that the image acquired in the image acquisition step is an image of the jellyfish detected by emitting sound waves into the water using sonar.

6. 6. A method for controlling the operation of a dust removal device based on a jellyfish attack prediction using the artificial intelligence according to claim 4 or 5, A method for controlling the operation of a dust removal device based on jellyfish attack predictions using artificial intelligence, characterized in that the computer uses artificial intelligence to recognize images in the jellyfish detection step, learn the operating status of the dust removal device, and the water level difference in the water level difference acquisition step, and change the setting values ​​in the operation control step.

Images