Damage detection device and damage detection program

The damage detection device and program improve the accuracy of detecting rotor blade damage by selecting and processing images from optimal angles or areas, using a supervised model to enhance visibility and detect issues like lightning strikes and erosion.

JP2024161841AInactive Publication Date: 2024-11-20MMGUARD INC
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
JP2023076933
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-08
Publication Date
2024-11-20
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing damage detection methods for wind turbine rotor blades are ineffective when the blades are photographed from certain orientations, leading to potential undetected damage due to obscured visibility.

Method used

A damage detection device and program that selects and processes images of wind turbine rotor blades from specific attitudes or areas to enhance detection accuracy, utilizing a supervised model trained on images from optimal viewing angles or areas, and detects damage based on temperature distributions or visible light analysis.

Benefits of technology

Enhances the accuracy of damage detection in rotor blades by ensuring images are captured from optimal viewing angles or areas, improving the detection of damage such as lightning strikes, erosion, and other surface issues.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024161841000001_ABST
    Figure 2024161841000001_ABST
Patent Text Reader

Abstract

To accurately detect damage to a rotor blade compared with the case of detecting damage to a rotor blade on the basis of an objective image which is acquired irrespective of a posture in which the image of the rotor blade was captured.SOLUTION: A damage detection device 10 has an acquisition unit 10A for acquiring, as an objective image, a captured image in which a posture of one rotor blade of a plurality of rotor blades is closest to a prescribed posture from a plurality of captured images of a windmill having a plurality of rotor blades that rotates around a common rotating shaft as captured from the side or in which an area of one rotor blade is closest to a prescribed area, and a detection unit 10B for detecting damage to the rotor blade on the basis of the acquired objective image.SELECTED DRAWING: Figure 4
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a damage detection device and a damage detection program. [Background technology]

[0002] Patent Document 1 discloses a report creation device. This report creation device includes an image capture means for capturing an image of a part of a wind power generation facility as an observation target by a capturing means attached to an unmanned aerial vehicle, and a flight position information capture means for capturing information on the flight position of the unmanned aerial vehicle. This report creation device also includes a data collection means for linking the image with the flight position information, a selection means for selecting an image including the observation target, and a calculation means for calculating the position information of the observation target based on the flight position information linked to the selected image. This report creation device also includes a report creation means for creating an inspection report including at least an image and the position information of the observation target included in the image, and an output means for outputting the inspection report.

[0003] Patent Document 2 discloses a method for inspecting a wind power generation facility. This inspection method includes a step of an unmanned aerial vehicle flying autonomously up to the sky above a wind turbine of the wind power generation facility, and a step of a controller provided in the unmanned aerial vehicle determining a current orientation of a nacelle of the wind turbine above the wind turbine. This inspection method also includes a step of the controller moving the unmanned aerial vehicle to a position directly facing the hub of the wind turbine based on the determined current orientation of the nacelle, and a step of the controller determining a current rotation angle of a first blade of the wind turbine at the position directly facing the hub. This inspection method also includes a step of the controller moving the unmanned aerial vehicle to an inspection start position of the first blade based on the determined current rotation angle of the first blade. This inspection method also includes a step of the controller taking an inspection image of the first blade with a camera mounted on the unmanned aerial vehicle while moving the unmanned aerial vehicle along the longitudinal direction of the first blade from the inspection start position. This inspection method also includes a step of automatically determining a damaged portion of the first blade based on the inspection image of the first blade.

[0004] US Patent No. 5,999,333 discloses a method for remotely inspecting a wind turbine blade in situ. The method includes imaging a portion of the wind turbine blade with a thermal imaging camera after sunset when the blade has reached substantial thermal equilibrium with the surrounding air to generate thermal imaging data. The method also includes analyzing the thermal imaging data to identify abnormal thermal patterns that may be indicative of frictional heating in or around defects in the wind turbine blade. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2018-181235 A [Patent Document 2] Patent Publication No. 2022-108420 [Patent Document 3] JP 2018-40807 A Summary of the Invention [Problem to be solved by the invention]

[0006] There is a technology that detects damage to a rotor blade of a wind turbine based on a target image obtained by photographing the rotor blade. However, with this technology, depending on the orientation of the rotor blade, the target image may not show damage, for example, when the rotor blade is photographed from above while in a perpendicular orientation to the horizontal plane. Therefore, even if a damaged rotor blade is photographed, this technology may not be able to detect the damage to the rotor blade.

[0007] The present disclosure aims to provide a damage detection device and a damage detection program that can detect damage to a rotor blade with high accuracy compared to detecting damage to a rotor blade based on a target image obtained regardless of the posture in which the rotor blade is photographed. [Means for solving the problem]

[0008] In order to achieve the above-mentioned object, the damage detection device of the first aspect has an acquisition unit that acquires, from a plurality of images obtained by photographing a wind turbine having a plurality of rotors rotating on a common rotation axis from the side, an image in which the attitude of one of the plurality of rotors is closest to a predetermined attitude, or an image in which the one rotor is closest to a predetermined area, as a target image, and a detection unit that detects damage to the rotor based on the acquired target image.

[0009] Further, in the damage detection device of the second aspect, in the damage detection device of the first aspect, the target image is an image acquired by inputting the plurality of captured images into a trained supervised model, and the supervised model is a model trained in advance using as teacher data a plurality of learning images acquired by photographing a wind turbine from the side, the image in which the attitude of one of the plurality of rotors of the wind turbine is closest to the predetermined attitude, or the image in which the area of ​​the one rotor is closest to the predetermined area.

[0010] In addition, in the damage detection device of the third aspect, in the damage detection device of the first or second aspect, the acquisition unit acquires as the target image from a plurality of images obtained by photographing the wind turbine from above, an image in which the attitude of one of the plurality of rotors is closest to an attitude parallel to a horizontal plane, or an image in which the area of ​​the one rotor is closest to a maximum area.

[0011] In addition, a damage detection device according to a fourth aspect is a damage detection device according to any one of the first to third aspects, wherein the detection unit detects damage to the rotor blade based on the temperature distribution of the acquired target image.

[0012] In addition, a damage detection program according to a fifth aspect causes a computer to execute a process of acquiring, from a plurality of images obtained by photographing a wind turbine having a plurality of rotors rotating on a common axis from the side, an image in which the attitude of one of the plurality of rotors is closest to a predetermined attitude, or an image in which the one rotor is closest to a predetermined area, as a target image, and detecting damage to the rotor based on the acquired target image. Effect of the Invention

[0013] According to the first or fifth aspect, damage to the rotor blade can be detected with high accuracy, compared to a case where damage to the rotor blade is detected based on a target image obtained regardless of the attitude of the rotor blade when it is photographed.

[0014] According to the second aspect, it is possible to acquire a target image with higher accuracy than when a plurality of captured images are not input to a supervised model.

[0015] According to the third aspect, damage to the rotor can be detected with high accuracy compared to a case where an image in which a rotor blade's attitude is closest to being parallel to the horizontal plane or an image in which a rotor blade's area is closest to being maximum is not acquired as the target image from among a plurality of captured images.

[0016] According to the fourth aspect, damage to the rotor blade can be detected with high accuracy, compared to a case where the temperature distribution of the target image is not used. [Brief description of the drawings]

[0017] [Figure 1] FIG. 2 is a schematic diagram illustrating an example of a hardware configuration of the damage detection system according to the embodiment. [Diagram 2] FIG. 1 is a block diagram showing an example of a configuration of a wind turbine according to an embodiment. [Diagram 3] 1 is a block diagram showing an example of a hardware configuration of a damage detection device according to an embodiment; [Figure 4] 2 is a block diagram showing an example of a functional configuration of a damage detection device according to an embodiment. FIG. [Diagram 5]10 is a flowchart showing an example of a flow of a damage detection process according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] Hereinafter, an example of an embodiment of the present disclosure will be described with reference to the drawings. In each drawing, the same or equivalent components and parts are given the same reference numerals. In addition, the dimensional ratios of the drawings are exaggerated for the convenience of explanation and may differ from the actual ratios.

[0019] As shown in FIG. 1, a damage detection system 100 according to this embodiment includes a damage detection device 10 and an aircraft 20.

[0020] The damage detection device 10 and the aircraft 20 are capable of communicating with each other via a communication means N. In this embodiment, a communication line such as the Internet, a LAN (Local Area Network), or a WAN (Wide Area Network) is applied as the communication means N. Also, in this embodiment, a wireless communication line is applied as the communication means N. However, this is not limited to this example. A wired communication line may be applied as the communication means N, or a combination of wired and wireless communication lines may be applied.

[0021] In this embodiment, a drone, which is a type of unmanned aircraft, is used as the aircraft 20. However, the present invention is not limited to this example. For example, the aircraft 20 may be a radio-controlled or manned aircraft.

[0022] As shown in FIG. 1, the aircraft 20 includes a camera 25 and a communication I / F (Interface) 27.

[0023] The camera 25 is an imaging device that captures images of the wind turbine 30. In this embodiment, an infrared camera capable of capturing infrared images is used as the camera 25. However, the present invention is not limited to this example. For example, a visible light camera capable of capturing visible light images may be used as the camera 25.

[0024] As shown in Fig. 2, the wind turbine 30 has a plurality of rotors 32A-32C that rotate on a common rotation shaft 31. In the following, when there is no need to distinguish between the plurality of rotors 32A-32C, they will simply be referred to as "rotor 32." Furthermore, the number of rotors 32 that the wind turbine 30 has is not limited to the example shown in Fig. 2.

[0025] Returning to FIG. 1, the communication I / F 27 is an interface for communicating with other devices such as the damage detection device 10.

[0026] In addition, the damage detection system 100 may use any device having an imaging device for imaging the wind turbine 30 instead of the flying object 20.

[0027] The damage detection device 10 is an information processing device such as a personal computer, a mobile terminal (for example, a smartphone, a mobile phone, or a tablet terminal), or a server.

[0028] 3, the damage detection device 10 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a display unit 16, and a communication I / F 17. Each component is connected to each other so as to be able to communicate with each other via a bus 19. The CPU 11 corresponds to a computer according to this embodiment.

[0029] The CPU 11 is a central processing unit, and executes various programs and controls each part. That is, the CPU 11 reads the programs from the ROM 12 and executes the programs using the RAM 13 as a working area. The CPU 11 controls each of the above components and performs various arithmetic processing according to the programs stored in the ROM 12. In this embodiment, a damage detection program 200 and a supervised model 202 are stored in the ROM 12.

[0030] The ROM 12 stores various programs and various data. The RAM 13 serves as a working area for temporarily storing programs or data.

[0031] The display unit 16 is, for example, a liquid crystal display, and displays various types of information.

[0032] The communication I / F 17 is an interface for communicating with other devices such as the flying object 20.

[0033] Next, a functional configuration of the damage detection device 10 according to this embodiment will be described with reference to Fig. 4. As shown in Fig. 4, the damage detection device 10 of this embodiment functions as an acquisition unit 10A, a detection unit 10B, and an output unit 10C by the CPU 11 executing a damage detection program 200 stored in the ROM 13.

[0034] The acquisition unit 10A acquires multiple captured images obtained by the flying object 20 photographing the wind turbine 30 from the side via the camera 25. In this embodiment, the side refers to any direction excluding the front of the wind turbine 30. In other words, the acquisition unit 10A acquires multiple captured images obtained by the camera 25 photographing the wind turbine 30 from at least one of the top, bottom, left, and right sides via the communication I / F 17. In the following, the multiple captured images of the wind turbine 30 photographed from the side acquired by the acquisition unit 10A are also simply referred to as "multiple captured images."

[0035] Then, the acquisition unit 10A acquires, from among the multiple captured images, a captured image in which the posture of one of the multiple rotors 32 is closest to a predetermined posture, or a captured image in which one rotor 32 is closest to a predetermined area, as a target image.

[0036] In this embodiment, damage that is likely to occur on the upper surface of the rotor 32, such as a lightning strike mark, is detected by the detection unit 10B, which will be described later. Therefore, the acquisition unit 10A acquires, from among the multiple captured images, multiple captured images obtained by capturing images of the wind turbine 30 from above. Then, the acquisition unit 10A acquires, from among the multiple captured images obtained by capturing images of the wind turbine 30 from above, an image in which the attitude of one rotor 32 is closest to an attitude parallel to the horizontal plane 40, or an image in which the area of ​​one rotor 32 is closest to the maximum area, as a target image.

[0037] In this embodiment, as the orientation parallel to the horizontal plane 40, an orientation in which at least one of the sides constituting the rotor 32 is parallel to the horizontal plane 40 is applied. For example, in the example shown in FIG. 2, the side constituting the rotor 32A indicated by the dashed line is in an orientation parallel to the horizontal plane 40. Therefore, in the example shown in FIG. 2, the acquisition unit 10A acquires, as the target image, a photographed image obtained by photographing the rotor 32A from above. However, this is not limited to this example. For example, as the orientation parallel to the horizontal plane 40, an orientation in which the perpendicular bisector of the shortest side or the longest side of the sides constituting the rotor 32, or the bisector of the smallest angle of the angles constituting the rotor 32, etc., is parallel to the horizontal plane 40 may be applied.

[0038] In addition, as long as the relative positional relationship between the rotor 32 and the camera 25 when the rotor 32 parallel to the horizontal plane 40 is photographed from above is maintained, the photographing direction may be any direction. For example, the acquisition unit 10A may acquire a plurality of photographed images obtained by photographing the wind turbine 30 from a side other than the top surface from among the plurality of photographed images. Then, the acquisition unit 10A may acquire, as the target image, a photographed image in which the attitude of one of the plurality of rotors 32 is closest to a predetermined attitude, or a photographed image in which one of the plurality of rotors 32 is closest to a predetermined area from among the plurality of photographed images obtained by photographing the wind turbine 30 from a side other than the top surface. In this case, the predetermined attitude is not limited to the above-mentioned attitude parallel to the horizontal plane 40, and may be a posture perpendicular to the horizontal plane 40, or a posture in which the angle with the horizontal plane 40 is irrelevant, etc. In addition, the predetermined area is not limited to the above-mentioned maximum area, and may be a specific area value of the rotor 32, or a minimum area in which the area of ​​one rotor 32 is the smallest, etc. In this case, the detection unit 10B described later can detect damage that is likely to occur on other surfaces besides the upper surface of the rotor blade 32, such as erosion, dents, peeling, corrosion, oxidation, melting, or cracks, in addition to lightning strike marks.

[0039] In this embodiment, the acquisition unit 10A acquires a target image by inputting a plurality of captured images to the supervised model 202 stored in the ROM 12. The supervised model 202 is a learned model that has been trained in advance using as teacher data a plurality of learning images obtained by photographing a wind turbine from the side, in which an image in which the attitude of one of the plurality of rotors of the wind turbine is closest to a predetermined attitude, or an image in which the area of ​​one rotor is closest to a predetermined area. In this embodiment, as the supervised model 202, a learned model that has been trained in advance using as teacher data a plurality of learning images obtained by photographing a wind turbine from above, in which an image in which the attitude of one rotor of the wind turbine is closest to an attitude parallel to the horizontal plane 40, or an image in which the area of ​​one rotor is closest to a maximum area, is applied. By inputting a plurality of captured images in which the rotor 32 is in various attitudes to the supervised model 202, a captured image in which the attitude of the rotor 32 is closest to a predetermined attitude, or an image in which the area of ​​one rotor 32 is closest to a predetermined area, is determined and output from among the plurality of captured images.

[0040] However, this is not a limitation. As the supervised model 202, a trained model may be applied that has been trained in advance using as training data a plurality of learning images obtained by photographing the wind turbine from a side other than the top, in which an image in which the attitude of one rotor of the wind turbine is closest to a predetermined attitude, or an image in which one rotor is closest to a predetermined area. In this case, the predetermined attitude is not limited to the attitude parallel to the horizontal plane 40 described above, but may be an attitude perpendicular to the horizontal plane, or an attitude in which the angle with the horizontal plane is irrelevant, etc. Furthermore, the predetermined area is not limited to the maximum area described above, but may be a specific area value of the rotor, or a minimum area in which the area of ​​one rotor is the smallest, etc.

[0041] The acquisition unit 10A may acquire the target image by any method. For example, when acquiring the attitude or area of ​​the rotor 32 photographed by the flying object 20 and transmitting these to the damage detection device 10 together with the photographed image, the acquisition unit 10A may acquire the target image from among the multiple photographed images based on the attitude or area of ​​the rotor 32 that has been transmitted.

[0042] The acquisition unit 10A may also acquire a target image directly without acquiring a plurality of captured images. For example, the acquisition unit 10A may acquire a target image acquired in advance by a device other than the damage detection device 10 via the communication I / F 17.

[0043] The detection unit 10B detects damage to the rotor 32 based on the target image acquired by the acquisition unit 10A. In this embodiment, an infrared camera is used as the camera 25, so the detection unit 10B detects damage to the rotor 32 based on the temperature distribution of the target image, which is an infrared image. Specifically, the detection unit 10B detects a portion in the target image where the temperature difference with the surroundings is equal to or greater than a predetermined temperature as a portion where damage has occurred. For example, the detection unit 10B detects damage to the rotor 32 by inputting the target image into a model that has been trained in advance using a pair of an image of the rotor 32 for learning and the presence or absence of damage to the rotor 32 for learning as teacher data.

[0044] When a visible light camera is used as the camera 25, the detection unit 10B may detect damage to the rotor 32 based on the target image, which is a visible light image. In this case, the detection unit 10B also detects damage to the rotor 32 using, for example, machine learning.

[0045] The output unit 10C outputs whether or not the detection unit 10B has detected damage to the rotor blade 32 (hereinafter referred to as "detection presence / absence"). In this embodiment, the output unit 10C outputs the detection presence / absence to the display unit 16. However, this is not limiting. For example, the output unit 10C may output the detection presence / absence to a device other than the damage detection device 10.

[0046] Next, the operation of the damage detection device 10 according to this embodiment will be described.

[0047] First, the damage detection process executed in the damage detection device 10 will be described with reference to Fig. 5. The damage detection process is executed by the CPU 11 functioning as an acquisition unit 10A, a detection unit 10B, and an output unit 10C. The damage detection process corresponds to a process that the damage detection program 200 causes the CPU 11 to execute.

[0048] In step S100 of FIG. 5, the CPU 11 acquires a plurality of photographed images obtained by the flying object 20 photographing the wind turbine 30 from the side via the camera 25.

[0049] In step S102, the CPU 11 acquires a target image. Specifically, the CPU 11 acquires the target image by inputting a plurality of captured images to the supervised model 202.

[0050] In step S104, the CPU 11 detects damage to the rotor blades 32. Specifically, the CPU 11 detects damage to the rotor blades 32 based on the temperature distribution of the target image.

[0051] In step S106, the CPU 11 outputs whether or not damage has been detected, and ends this damage detection process.

[0052] Although the embodiments have been described above, the technical scope of the present invention is not limited to the scope described in the above embodiments. Various modifications or improvements can be made to the above embodiments without departing from the gist of the invention, and such modifications or improvements are also included in the technical scope of the present invention.

[0053] Furthermore, the above-mentioned embodiments do not limit the invention according to the claims, and not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention. The above-mentioned embodiments include inventions at various stages, and various inventions are extracted by combining the multiple components disclosed. Even if some components are deleted from all the components shown in the embodiments, as long as the effect can be obtained, the configuration from which these some components are deleted can be extracted as an invention.

[0054] For example, in the above embodiment, the damage detection device 10 and the flying object 20 are configured as separate entities. However, this is not limited to this example. For example, the damage detection device 10 may be built into the flying object 20.

[0055] In the present embodiment, the damage detection device 10 and the flying object 20 are capable of communicating with each other via the communication means N. However, this is not limited to this example. For example, the damage detection device 10 and the flying object 20 do not have to be capable of communicating with each other via the communication means N. In this case, for example, the flying object 20 stores the images captured by the camera 25 in a storage device such as a memory card. Then, the CPU 11 of the damage detection device 10 acquires the images captured by the camera 25 from the storage device via an input / output I / F provided in the damage detection device 10.

[0056] In addition, in the above embodiments, the term "processor" refers to a processor in a broad sense, and includes general-purpose processors (e.g., CPU: Central Processing Unit, etc.) and dedicated processors (e.g., GPU: Graphics Processing Unit, ASIC: Application Specific Integrated Circuit, FPGA: Field Programmable Gate Array, programmable logic device, etc.).

[0057] In addition, the operations of the processor in the above embodiments may be performed not only by one processor, but also by multiple processors located at physically separate locations working together. The order of the operations of the processor is not limited to the order described in the above embodiments, and may be changed as appropriate.

[0058] In the present embodiment, the program is installed in the ROM, but the present invention is not limited to this. The program according to the present embodiment may be provided in a form recorded in a computer-readable storage medium. For example, the program according to the present embodiment may be provided in a form recorded in an optical disc such as a CD (Compact Disc)-ROM or a DVD (Digital Versatile Disc)-ROM, or in a semiconductor memory such as a USB (Universal Serial Bus) memory or a memory card. The damage detection program 200 according to the present embodiment may be obtained from an external device via the communication I / F 17.

[0059] In the above embodiment, the processing in the damage detection device 10 is implemented by a software configuration using a computer by executing a program, but the present invention is not limited to this. For example, the processing in the damage detection device 10 may be implemented by a hardware configuration or a combination of a hardware configuration and a software configuration.

[0060] In addition, the configurations of the damage detection device 10 and the aircraft 20 described in the above embodiment are merely examples, and it goes without saying that unnecessary parts may be deleted or new parts may be added within the scope that does not deviate from the gist of the present invention.

[0061] Furthermore, the processing flow in the damage detection device 10 described in the above embodiment is also one example, and it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged within the scope of the present invention.

[0062] Furthermore, the present disclosure is not limited to the above, and it goes without saying that various modifications can be made without departing from the spirit and scope of the present disclosure. [Explanation of symbols]

[0063] 10 Damage detection device 11 CPU 12 ROM 13 RAM 16 Display 17 Communication I / F 19 Bus 20 Flying Objects 25 Camera 27 Communication I / F 30 windmill 31 Rotation axis 32 Rotor 40 horizontal plane 100 Damage Detection System 200 Damage Detection Program 202 Supervised Model

Claims

1. an acquisition unit that acquires, from among a plurality of photographed images obtained by photographing a wind turbine having a plurality of rotors rotating on a common rotation axis from the side, a photographed image in which a posture of one of the plurality of rotors is closest to a predetermined posture, or a photographed image in which the one rotor is closest to a predetermined area, as a target image; a detection unit that detects damage to the rotor blade based on the acquired target image; A damage detection device comprising:

2. The target image is an image obtained by inputting the plurality of captured images into a trained supervised model, The supervised model is a model that has been trained in advance using as training data a plurality of learning images obtained by photographing a wind turbine from the side, the image in which the attitude of one of a plurality of rotors of the wind turbine is closest to the predetermined attitude, or the image in which the area of ​​the one rotor is closest to the predetermined area. The damage detection device of claim 1 .

3. The acquisition unit acquires, from among a plurality of captured images obtained by photographing the wind turbine from above, an image in which a posture of one of the plurality of rotors is closest to a posture parallel to a horizontal plane, or an image in which an area of ​​the one rotor is closest to a maximum area, as the target image. The damage detection device according to claim 1 or 2.

4. The detection unit detects damage to the rotor blade based on a temperature distribution in the acquired target image. The damage detection device according to claim 1 or 2.

5. Among a plurality of photographed images obtained by photographing a wind turbine having a plurality of rotors rotating on a common rotation axis from the side, a photographed image in which a posture of one of the plurality of rotors is closest to a predetermined posture or a photographed image in which the one rotor is closest to a predetermined area is obtained as a target image; Detecting damage to the rotor blade based on the acquired target image. Have a computer carry out the process, Damage detection program.

Citation Information

Patent Citations

  • System and method for ground-based inspection of wind turbine blade

    JP2018040807A

  • Report generator, wind power generation equipment inspection system, program, and method for generating inspection report of wind power generation equipment

    JP2018181235A

  • Inspection method for wind turbine generator facility

    JP2022108420A