Image recognition-based precise fault identification system and method for hydroelectric generator

Abstract

The application discloses an image recognition-based precise fault discrimination system and method for a hydro-generator, wherein a first image recognition component of the discrimination system is installed on the upper surface of the inner top cover of the hydro-generator unit, a shear pin signal device is installed on the back of the guide vane crank arm, a second image recognition component is installed on the upper rack of the hydro-generator unit, a damper sensor is installed above and below the damper shutter movement, an image processing component performs image processing, and a display screen displays the discrimination result; compared with artificial field judgment, the discrimination system can greatly improve the efficiency and realize real-time monitoring by acquiring image information of the angle between the guide vane crank arm and the connecting rod and distance information of the damper shutter and the lower surface of the brake ring; the discrimination method can make a judgment on the actual operation of the hydro-generator by combining the state of the shear pin signal device and the damper sensor with the result of the image processing component, realize precise judgment, and improve the fault handling efficiency.

Description

Technical Field

[0001] This invention relates to the field of hydropower production and operation technology, and in particular to a system and method for accurate fault diagnosis of hydro-generators based on image recognition. Background Technology

[0002] The shear pin in a large hydropower station is one of the most common guide vane drive protection devices in the guide vane mechanism of a hydro-generator unit. It plays a crucial role in the safe operation of the turbine. When foreign objects get stuck between the turbine guide vanes, the guide vanes are obstructed during opening and closing. At this time, the operating torque is transmitted to the guide vane through the shear pin. When the torque exceeds the shear force set by the shear pin, the shear pin of that guide vane is sheared off and disengaged from the control loop, thus protecting the guide vane from damage. Other guide vanes can still operate normally with the control loop, preventing the unit from going out of control. The airlock (usually referring to the air braking system) in a large hydropower station is an important auxiliary device for the hydro-generator unit, mainly used for the safe shutdown and operation protection of the unit.

[0003] In current hydropower station operation and maintenance practices, the status monitoring of shear pins and wind dampers relies on sensors for fault detection. After a sensor signal is lost (i.e., after a fault signal is issued), staff need to conduct on-site inspection and confirmation before taking any action. However, in actual hydropower station operation and maintenance, both shear pins and wind dampers have a certain false alarm rate for sensor signals. When troubleshooting, it's crucial to first rule out the possibility of false alarms and then determine the fault type. Furthermore, each hydropower unit has dozens of shear pins and wind dampers. The shear pins are located in the turbine room, and the wind dampers are located in the wind tunnel. The turbine room and wind tunnel are very confined spaces, and their distance from the central control room is considerable. On-site confirmation by staff takes a significant amount of time, delaying fault handling, affecting efficiency, and further impacting the unit's start-up success rate. Moreover, insufficient staff experience may lead to inaccurate confirmation.

[0004] Machine vision inspection is a technology that simulates human vision, thinking, and operation to inspect product quality. It utilizes high-precision optical imaging systems, powerful computer processing capabilities, and automated actuators to automatically detect various issues such as appearance, dimensions, and defects in products, comprehensively improving inspection efficiency and accuracy. In the actual operation and maintenance of hydropower stations, in the fault diagnosis of turbine generator units, the state of shear pins can be determined by observing the angle between the guide vane crank arm and the connecting rod on-site. Similarly, the lifting and lowering position of the wind gate can be determined by observing the distance between the upper surface of the wind gate and the lower surface of the turbine generator brake ring. Applying machine vision monitoring to the fault diagnosis of turbine generator units can yield accurate fault diagnosis results. Summary of the Invention

[0005] To overcome the above problems, the purpose of this invention is to provide a system and method for accurate fault diagnosis of hydro-generators based on image recognition. This system acquires image information of the angle between the guide vane crank arm and the connecting rod of the hydro-generator unit through a first image recognition component, acquires distance information between the air damper and the lower surface of the brake ring through a second image recognition component, and acquires specific parameter values ​​through an image processing component. Compared with manual on-site judgment, this system can greatly improve efficiency, achieve real-time monitoring, and thus promptly detect and handle faults. The diagnosis method combines the status of the shear pin signal device and the air damper sensor with the results of the image processing component to judge the actual operating condition of the hydro-generator, and outputs the judgment results on the display for easy observation and improved fault handling efficiency.

[0006] The technical solution adopted in this invention is:

[0007] A precise fault diagnosis system for hydro-generators based on image recognition includes a first image recognition component, a shear pin signal device, a second image recognition component, a damper sensor, an image processing component, and a display screen. The first image recognition component is installed on the upper surface of the inner top cover of the hydro-generator unit. The shear pin signal device is installed on the back of the guide vane crank arm of the hydro-generator unit. The second image recognition component is installed on the upper frame of the hydro-generator unit. The damper sensor includes an upper sensor and a lower sensor, which are respectively installed above and below the damper gate during movement. The image processing component receives image information from the first and second image recognition components and performs image processing. The display screen displays the final diagnosis result.

[0008] The first image recognition component includes a ring track, a motion bracket, a first industrial camera, and a first industrial lens. The ring track is fixedly installed on the upper surface of the inner top cover, the motion bracket is slidably connected to the ring track, the first industrial camera is fixedly installed on the motion bracket, and the first industrial lens is connected to the first industrial camera.

[0009] The second image recognition component comprises multiple sets, including a base, a bracket, a second industrial camera, and a second industrial lens. The base is fixedly installed on the upper surface edge of the upper frame near the air damper. The bracket is located on the upper surface of the base. The second industrial camera is fixedly connected to the bracket, and the second industrial lens is connected to the second industrial camera.

[0010] As a further description of the present invention, the outer side of the annular track is designed with a groove, and gear columns are evenly distributed in the groove on the outer side of the annular track. The motion support includes a sliding plate, a motor, gears, a camera mounting bracket, and pulleys. The sliding plate is slidably connected to the upper surface of the annular track. A groove is opened on one side above the sliding plate, and the motor is installed in the groove. The output shaft of the motor passes through the sliding plate and connects to the gear. The gear is adapted to the gear column. The camera mounting bracket is fixedly connected to the upper surface of the sliding plate and is located to the right of the motor. There are four pulleys, which are fixedly connected to the front and rear sides of the sliding plate respectively. The two pulleys on the groove side of the annular track are located on the left and right sides of the gear respectively, and the two pulleys on the other side are tangent to the inner side of the annular track respectively.

[0011] As a further description of the present invention, the second image recognition component also includes a visual light source and a light source controller. The bracket has an S-shaped structure. The second industrial camera is fixedly mounted on the upper surface of the bracket. The visual light source is mounted below the second industrial camera. The light source controller is connected to the visual light source and is located to the right of the visual light source.

[0012] The visual light source is a strip light source.

[0013] As a further description of the present invention, the airlock sensor is a pressure sensor.

[0014] As a further description of the present invention, the first image recognition component further includes a telescopic support column, wherein there are multiple telescopic support columns, which are installed at the position between the lower surface of the annular track and the inner top cover.

[0015] The base of the second image recognition component is a telescopic base.

[0016] As a further description of the present invention, the number of the second image recognition components is the same as the number of the wind gate plates of the hydro turbine generator, and the second image recognition components are installed at the location of each group of wind gate plates.

[0017] The fault identification method of the above-mentioned image recognition-based hydro-generator fault accuracy identification system includes the following steps:

[0018] S01: Data acquisition, obtaining the status of the shear pin, the status of the damper sensor, the image of the first image recognition component, and the image of the second image recognition component;

[0019] S02: Data processing,

[0020] The image captured by the first image recognition component is used in the image processing component to calculate the angle between the guide vane crank arm and the connecting rod. ,

[0021] The distance between the upper surface of the windshield gate and the lower surface of the turbine generator brake ring is calculated in the image processing component using the image captured by the second image recognition component. ;

[0022] S03: Fault diagnosis;

[0023] S04: Output the results on the display screen.

[0024] As a further description of the present invention, the determination process of S03 is as follows:

[0025] When the image features are detected When the angle is less than or equal to the preset angle threshold, it is determined that the feature is consistent with the feature in the preset image feature library of the normal position state of the shear pin.

[0026] When the image features are detected When the angle exceeds the preset threshold, it is determined that the feature comparison with the preset image feature library of the normal position state of the shear pin is inconsistent.

[0027] When the image features are detected When the distance is less than or equal to the preset lifting distance threshold, it is determined that the feature is consistent with the feature in the preset windshield gate lifting position state image feature library;

[0028] When the image features are detected If the distance exceeds the preset threshold, it is determined that the feature comparison with the preset windshield gate position status image feature library is inconsistent.

[0029] When the image features are detected When the distance is less than or equal to the preset falling distance threshold, it is determined that the feature is consistent with the feature in the preset air damper gate falling position state image feature library;

[0030] When the image features are detected If the distance exceeds the preset threshold, it is determined that the feature is inconsistent with the feature in the preset image feature library of the windshield gate's falling position.

[0031] As a further description of the present invention, the angle threshold is 17°, the lifting distance threshold is 35mm, and the falling distance threshold is 5mm.

[0032] As a further description of the present invention, the output process of S04 is as follows:

[0033] If the shear pin sensor status is sheared, the image features will show... If the feature does not match the preset image feature library, the output will be: Cut off by the X-shaped cutting pin.

[0034] If the cut pin is in an uncut state, the image features... If the feature does not match the feature in the preset image feature library, the output will be: "X-shaped cutting pin may have been cut, but no signal is reported at this time."

[0035] If the cut pin is in the cut state, the image features... When the feature matches the feature in the preset image feature library, output: false alarm of cutting signal of cutting pin X;

[0036] If the cut pin is in an uncut state, the image features... When the feature matches the feature in the preset image feature library, the output is: "X-numbered shear pin is in normal condition."

[0037] If the air damper sensor is in the "lifted" state, meaning the upper sensor detects a pressure signal and the air damper plate position matches the preset air damper plate lift position status image feature library, the output will be: "Air damper X is in normal lift status."

[0038] If the damper sensor is in the "lifted" state, meaning the upper sensor detects a pressure signal and the damper gate position does not match the preset damper gate lift position image feature library, the output will be: "Damper X is in an abnormal state."

[0039] If the air damper sensor is in a falling state, that is, the lower sensor detects a pressure signal and the position of the air damper plate matches the preset air damper plate falling position state image feature library, the output is: Air damper X is in a normal falling state.

[0040] If the air damper sensor is in a falling state, that is, the lower sensor detects a pressure signal and the position of the air damper plate does not match the preset air damper plate falling position state image feature library, the output will be: Air damper No. X is in an abnormal state.

[0041] If the status of the air damper sensor is not up or down, that is, neither the upper nor the lower sensor detects a pressure signal, and the position of the air damper plate is inconsistent with the preset air damper plate up or down position status image feature library, the output will be: Air damper No. X position is abnormal.

[0042] If the status of the air damper sensor is not either raised or lowered, that is, neither the upper nor lower sensor detects a pressure signal, and the position of the air damper plate does not match the preset air damper plate middle position status image feature library, the output will be: Air damper No. X is in an abnormal state.

[0043] The beneficial effects of this invention are:

[0044] This invention relates to a precise fault diagnosis system for hydro-generators based on image recognition. The system includes a first image recognition component, a shear pin signal device, a second image recognition component, a damper sensor, an image processing component, and a display screen. The system acquires image information of the angle between the guide vane crank arm and the connecting rod of the hydro-generator unit through the first image recognition component, acquires distance information between the damper plate and the lower surface of the brake ring through the second image recognition component, and acquires specific parameter values ​​through the image processing component. Compared to manual on-site judgment, this system significantly improves efficiency, enables real-time monitoring, and allows for timely fault detection and handling, thereby increasing the first-time start-up success rate of the hydro-generator unit.

[0045] This invention relates to an image recognition-based system for accurate fault diagnosis of hydro-generators. The first image recognition component includes a circular track, a moving support, a first industrial camera, and a first industrial lens. The second image recognition component comprises multiple sets, including a base, a support, a second industrial camera, and a second industrial lens. This system, combined with the actual fault conditions of hydro-generator sets, uses the moving first image recognition component for shear pin monitoring and multiple stationary second image recognition components for wind gate monitoring. Since the failure rate of shear pins is low and the failure rate of wind gates is high, this system design achieves real-time monitoring and accurate diagnosis while ensuring the lowest system cost and achieving optimal system configuration.

[0046] This invention relates to an image recognition-based system for accurate fault diagnosis of hydro-generators. The second image recognition component further includes a visual light source and a light source controller. Due to the low brightness inside the wind tunnel and the high precision required for monitoring the rising and falling position of the wind gate, which needs to reach 1mm, a visual light source is set in the second image recognition component. The visual light source adopts a strip light source, which has high brightness and uniform illumination, providing a high-brightness and uniform lighting effect. This makes every detail in the image recognition site clearly visible, which helps to improve detection accuracy and production efficiency, further ensuring the accuracy of the acquired wind gate image information, improving the system's judgment accuracy, and achieving precise judgment.

[0047] This invention uses a fault diagnosis method for a hydro-generator fault accuracy system based on image recognition. The diagnosis method combines the status of the shear pin signal device and the air damper sensor with the results of the image processing component to judge the actual operating status of the hydro-generator, and outputs the judgment results on the display for easy observation and improved fault handling efficiency. Attached Figure Description

[0048] Figure 1 This is a block diagram of the overall structure of the image recognition-based hydro-generator fault accurate identification system proposed in this invention;

[0049] Figure 2This is a three-dimensional structural diagram of the first image recognition component of the image recognition-based hydro-generator fault accurate identification system proposed in this invention.

[0050] Figure 3 This is a schematic diagram of the moving support structure of the image recognition-based hydro-generator fault accurate identification system proposed in this invention;

[0051] Figure 4 This is a schematic diagram of the telescopic support column structure of the image recognition-based hydro-generator fault accurate identification system proposed in this invention;

[0052] Figure 5 This is a three-dimensional structural diagram of the second image recognition component of the image recognition-based hydro-generator fault accurate identification system proposed in this invention;

[0053] Figure 6 This is a schematic diagram of the visual light source and light source controller structure of the image recognition-based hydro-generator fault accurate identification system proposed in this invention.

[0054] Figure 7 This is a flowchart of the fault identification method for a hydro-generator fault accuracy identification system based on image recognition proposed in this invention.

[0055] Explanation of reference numerals in the attached figures

[0056] 1-First image recognition component,

[0057] 11-Circular track, 111-Gear column,

[0058] 12-Motion bracket, 121-Sliding plate, 122-Motor, 123-Gear, 124-Camera mount, 125-Pulley,

[0059] 13 - First Industrial Camera

[0060] 14-First Industrial Lens

[0061] 15- Telescopic support column,

[0062] 2-Shear pin signal device,

[0063] 3-Second image recognition component,

[0064] 31-Base, 32-Bracket, 33-Second industrial camera, 34-Second industrial lens, 35-Vision light source, 36-Light source controller

[0065] 4-Air damper sensor,

[0066] 41 - Upper sensor, 42 - Lower sensor

[0067] 5-Image processing components,

[0068] 6- Display screen. Detailed Implementation

[0069] The specific embodiments of the present invention are described below with reference to the accompanying drawings and examples:

[0070] It should be noted that the structures, proportions, sizes, etc. illustrated in the accompanying drawings of this specification are only used to complement the content disclosed in the specification, so that those skilled in the art can understand and read them, and are not intended to limit the conditions under which the present invention can be implemented. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0071] Furthermore, the terms such as "upper," "lower," "left," "right," "middle," and "one" used in this specification are merely for clarity of description and are not intended to limit the scope of the invention. Any changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

[0072] like Figures 1-7 As shown, it illustrates a specific embodiment of the present invention:

[0073] Example 1

[0074] The image recognition-based hydro-generator fault accurate identification system includes a first image recognition component 1, a shear pin signal device 2, a second image recognition component 3, a damper sensor 4, an image processing component 5, and a display screen 6. The first image recognition component 1 is installed on the upper surface of the inner top cover of the hydro-generator unit. The shear pin signal device 2 is installed on the back of the guide vane crank arm of the hydro-generator unit. The second image recognition component 3 is installed on the upper frame of the hydro-generator unit. The damper sensor 4 includes an upper sensor 41 and a lower sensor 42, which are respectively installed above and below the damper plate during movement. The image processing component 5 receives image information from the first image recognition component 1 and the second image recognition component 3 and performs image processing. The display screen 6 displays the final identification result.

[0075] In this embodiment, as Figure 1 As shown, the discrimination system obtains image information of the angle between the guide vane crank arm and the connecting rod of the hydro-generator unit through the first image recognition component 1, obtains distance information between the wind gate plate and the lower surface of the brake ring through the second image recognition component 3, and processes the image information of the first image recognition component 1 and the second image recognition component 3 through the image processing component 5 to obtain specific parameter values. Compared with manual on-site judgment, this discrimination system can greatly improve efficiency, realize real-time monitoring, and thus promptly detect and handle faults, thereby improving the first-time start-up success rate of the hydro-generator unit.

[0076] The first image recognition component 1 includes an annular track 11, a motion bracket 12, a first industrial camera 13, and a first industrial lens 14. The annular track 11 is fixedly installed on the upper surface of the inner top cover. The motion bracket 12 is slidably connected to the annular track 11. The first industrial camera 13 is fixedly installed on the motion bracket 12. The first industrial lens 14 is connected to the first industrial camera 13.

[0077] In this embodiment, as Figure 2 As shown, the motion support 12 can run on the circular track 11. The movement of the motion support 12 realizes the position movement of the first industrial camera 13 on the motion support 12, thereby acquiring image information of the angle between the guide vane crank arm and the connecting rod above the guide vane at different positions. In actual use, the position of the guide vane is confirmed according to the positioning point of the motion support 12, thereby clarifying which guide vane the image information collected at that position is the image information above. The first industrial camera 13 can also be used to shoot video, extract image information from the video, and then obtain feature information.

[0078] The second image recognition component 3 has multiple sets, including a base 31, a bracket 32, a second industrial camera 33, and a second industrial lens 34. The base 31 is fixedly installed on the upper surface edge of the upper frame near the air damper plate. The bracket 32 ​​is located on the upper surface of the base 31. The second industrial camera 33 is fixedly connected to the bracket 32. The second industrial lens 34 is connected to the second industrial camera 33.

[0079] Specifically, the number of the second image recognition components 3 is the same as the number of the wind gate panels of the hydro turbine generator, and the second image recognition components 3 are installed at the location of each group of wind gate panels.

[0080] In this embodiment, as Figure 5 As shown, the second image recognition component 3 is used to acquire the image of the windshield gate, and is implemented using a stationary support 32.

[0081] In this embodiment, the system combines the actual fault conditions of the hydro-generator unit. It uses a moving first image recognition component 1 for shear pin monitoring and multiple static second image recognition components 3 for wind gate monitoring. Since the failure rate of shear pin is low and the failure rate of wind gate is high, this system design can achieve real-time monitoring and accurate identification while ensuring the lowest system cost, achieving optimal system configuration, and maximizing energy efficiency.

[0082] Specifically, the outer side of the annular track 11 has a groove design, and gear columns 111 are evenly distributed in the groove on the outer side of the annular track 11. The motion support 12 includes a sliding plate 121, a motor 122, a gear 123, a camera mounting bracket 124, and pulleys 125. The sliding plate 121 is slidably connected to the upper surface of the annular track 11. A groove is opened on one side of the upper part of the sliding plate 121. The motor 122 is installed in the groove. The output shaft of the motor 122 passes through the sliding plate 121 and connects to the gear 123. The gear 123 is adapted to the gear column 111. The camera mounting bracket 124 is fixedly connected to the upper surface of the sliding plate 121 and is located to the right of the motor 122. There are four pulleys 125, which are fixedly connected to the front and rear sides of the sliding plate 121 respectively. The two pulleys 125 on the groove side of the annular track 11 are located on the left and right sides of the gear 123 respectively. The two pulleys 125 on the other side are tangent to the inner side of the annular track 11.

[0083] In this embodiment, as Figure 3 As shown, the motion support 12 moves on the annular track 11 as follows: the motor 122 rotates under the action of electrical energy, and the gear 123 connected to the output shaft of the motor 122 rotates. Since the gear 123 is adapted to the gear column 111 in the groove of the annular track 11, it moves between the gear columns 11 under the action of the rotation of the gear 123, so that the sliding plate 121 begins to slide on the upper surface of the annular track 11 and realizes circular motion around the annular track 11. The camera mounting bracket 124 is used to install the first industrial camera 13 to obtain image information from different positions. The pulley 125 is located on the front and rear sides of the sliding plate 121 to ensure that the sliding plate 121 moves between the annular track 11 and realizes the reliable movement of the motion support 12.

[0084] Example 2

[0085] Based on the above embodiments, in order to accurately determine the fault status of the hydro-generator unit, precise data information is required. Therefore, this second embodiment is proposed.

[0086] Specifically, the second image recognition component 3 also includes a visual light source 35 and a light source controller 36. The bracket 32 ​​has an S-shaped structure. The second industrial camera 33 is fixedly installed on the upper surface of the bracket 32. The visual light source 35 is installed below the second industrial camera 33. The light source controller 36 is connected to the visual light source 35 and is located on the right side of the visual light source 35.

[0087] The visual light source 35 is a strip light source.

[0088] Specifically, the airlock sensor 4 is a pressure sensor.

[0089] In this embodiment, as Figure 6 As shown, the second image recognition component 3 also includes a visual light source 35 and a light source controller 36. Since the windshield is installed inside the wind tunnel, where the brightness is low, and the monitoring of the rising and falling position of the windshield requires high precision (1mm), a visual light source 35 is set in the second image recognition component 3. The visual light source 35 is a strip light source, which has high brightness and uniform illumination, providing a high-brightness and uniform lighting effect. This makes every detail in the image recognition site clearly visible, which helps to improve detection accuracy and production efficiency, further ensuring the accuracy of the acquired windshield image information, improving the system's discrimination accuracy, and achieving precise judgment.

[0090] Example 3

[0091] Based on the above embodiments, since the guide vane crank arm and connecting rod have different height positions and the wind gate plate height positions of different specifications of hydro turbine generator sets, this embodiment is proposed to make the system more widely applicable.

[0092] Specifically, the first image recognition component 1 also includes a telescopic support column 15, and there are multiple telescopic support columns 15, which are installed between the lower surface of the annular track 11 and the inner top cover.

[0093] The base 31 of the second image recognition component 3 is a telescopic base.

[0094] In this embodiment, as Figure 4 As shown, a telescopic support column 15 is set below the annular track 11 of the first image recognition component 1, and the base 31 of the second image recognition component 3 is designed as a telescopic base. This makes it convenient for operators to adjust the height according to actual use needs and adapt to different height requirements. The telescopic design can be either the existing hydraulic telescopic or the existing snap-on telescopic, and the height adjustment accuracy is selected as 2mm.

[0095] Example 4

[0096] The method for accurately identifying faults in hydro-generators based on image recognition, as described above, includes the following steps:

[0097] S01: Data acquisition, acquiring the status of the shear pin in the shear pin signal device 2, the status of the air damper sensor 4, the image of the first image recognition component 1, and the image of the second image recognition component 3;

[0098] S02: Data processing,

[0099] The image captured by the first image recognition component 1 is used in the image processing component 5 to calculate the angle between the guide vane crank arm and the connecting rod. ,

[0100] The distance between the upper surface of the windshield gate and the lower surface of the turbine generator brake ring is calculated in the image processing component 5 using the image captured by the second image recognition component 3. ;

[0101] S03: Fault diagnosis;

[0102] S04: Output the results on display screen 6.

[0103] In this embodiment, as Figure 7 As shown, this discrimination method combines the status of the shear pin signal device 2 and the air damper sensor 4 with the results of the image processing component 5 to judge the actual operation of the hydro-generator, and outputs the judgment results on the display 6 for easy observation and improved fault handling efficiency.

[0104] Specifically, the determination process for S03 is as follows:

[0105] When the image features are detected When the angle is less than or equal to the preset angle threshold, it is determined that the feature is consistent with the feature in the preset image feature library of the normal position state of the shear pin.

[0106] When the image features are detected When the angle exceeds the preset threshold, it is determined that the feature comparison with the preset image feature library of the normal position state of the shear pin is inconsistent.

[0107] When the image features are detected When the distance is less than or equal to the preset lifting distance threshold, it is determined that the feature is consistent with the feature in the preset windshield gate lifting position state image feature library;

[0108] When the image features are detected If the distance exceeds the preset threshold, it is determined that the feature comparison with the preset windshield gate position status image feature library is inconsistent.

[0109] When the image features are detected When the distance is less than or equal to the preset falling distance threshold, it is determined that the feature is consistent with the feature in the preset air damper gate falling position state image feature library;

[0110] When the image features are detected If the distance exceeds the preset threshold, it is determined that the feature is inconsistent with the feature in the preset image feature library of the windshield gate's falling position.

[0111] In this embodiment, the judgment process is divided into two aspects: one is to determine the angle between the guide vane crank arm and the connecting rod. The angle is compared with the actual preset angle threshold. If the included angle value is not within the threshold range, it is determined that it is inconsistent with the feature comparison in the feature library. On the other hand, the distance between the upper surface of the wind gate and the lower surface of the turbine generator brake ring is also considered. The results are compared with preset lifting distance thresholds and falling distance thresholds to determine whether the comparison results are consistent or inconsistent.

[0112] Specifically, the output process of S04 is as follows:

[0113] If the cut pin is in the cut state, the image features... If the feature does not match the preset image feature library, the output will be: Cut off by the X-shaped cutting pin.

[0114] If the cut pin is in an uncut state, the image features... If the feature does not match the feature in the preset image feature library, the output will be: "X-shaped cutting pin may have been cut, but no signal is reported at this time."

[0115] If the cut pin is in the cut state, the image features... When the feature matches the feature in the preset image feature library, output: false alarm of cutting signal of cutting pin X;

[0116] If the cut pin is in an uncut state, the image features... When the feature matches the feature in the preset image feature library, the output is: "X-numbered shear pin is in normal condition."

[0117] If the state of the air damper sensor 4 is "lifted", that is, the upper sensor 41 detects a pressure signal and the position of the air damper plate matches the preset air damper plate lift position state image feature library, the output is: "Air damper X is in normal lift state".

[0118] If the state of the air damper sensor 4 is "lifted", that is, the upper sensor 41 detects a pressure signal and the position of the air damper plate does not match the preset air damper plate lift position state image feature library, the output will be: "Air damper No. X is in an abnormal state".

[0119] If the state of the air damper sensor 4 is falling, that is, the lower sensor 42 detects a pressure signal and the position of the air damper plate matches the preset air damper plate falling position state image feature library, the output is: the falling state of air damper X is normal.

[0120] If the state of the air damper sensor 4 is falling, that is, the lower sensor 42 detects a pressure signal, and the position of the air damper plate is inconsistent with the preset air damper plate falling position state image feature library, the output is: Air damper No. X is in an abnormal state.

[0121] If the state of the air damper sensor 4 is not up or down, that is, neither the upper sensor 41 nor the lower sensor 42 detects a pressure signal, and the position of the air damper plate is inconsistent with the preset air damper plate up or down position state image feature library, the output will be: Air damper No. X position is abnormal.

[0122] If the state of the air damper sensor 4 is neither raised nor lowered, that is, neither the upper sensor 41 nor the lower sensor 42 detects a pressure signal, and the position of the air damper plate is inconsistent with the preset air damper plate middle position state image feature library, the output will be: Air damper No. X is in an abnormal state.

[0123] In this embodiment, the output results are combined with the status of the shear pin and the status of the wind gate sensor to ensure the accuracy of the fault diagnosis of the turbine generator guide vane, so as to take corresponding measures in a timely manner, improve the efficiency of fault handling, and directly output the results on the display screen 6 for easy observation.

[0124] Example 5

[0125] Specifically, the angle threshold is 17°, the lifting distance threshold is 35mm, and the falling distance threshold is 5mm.

[0126] In this embodiment, the threshold range is obtained based on the actual experience and test results of hydropower plants, which enables accurate judgment of turbine generator unit faults.

[0127] The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

[0128] Many other changes and modifications can be made without departing from the concept and scope of this invention. It should be understood that this invention is not limited to the specific embodiments, and the scope of this invention is defined by the appended claims.

Claims

1. A precise fault diagnosis system for hydro-generators based on image recognition, characterized in that, The system includes a first image recognition component (1), a shear pin signal device (2), a second image recognition component (3), a damper sensor (4), an image processing component (5), and a display screen (6). The first image recognition component (1) is installed on the upper surface of the inner top cover of the turbine generator set. The shear pin signal device (2) is installed on the back of the guide vane crank arm of the turbine generator set. The second image recognition component (3) is installed on the upper frame of the turbine generator set. The damper sensor (4) is a pressure sensor, including an upper sensor (41) and a lower sensor (42), which are installed above and below the damper gate plate, respectively. The image processing component (5) receives the image information from the first image recognition component (1) and the second image recognition component (3) and performs image processing: combining the shear pin status obtained from the shear pin signal device (2) and the damper sensor status obtained from the damper sensor (4), and making a comprehensive judgment with the result obtained by processing the image information, the specific fault situation is determined. The display screen (6) displays the final judgment result. The first image recognition component (1) includes a ring track (11), a motion bracket (12), a first industrial camera (13), and a first industrial lens (14). The ring track (11) is fixedly installed on the upper surface of the inner top cover. The motion bracket (12) is slidably connected to the ring track (11). The first industrial camera (13) is fixedly installed on the motion bracket (12). The first industrial lens (14) is connected to the first industrial camera (13). It also includes a telescopic support column (15). There are multiple telescopic support columns (15), which are installed between the lower surface of the ring track (11) and the inner top cover. The second image recognition component (3) has multiple sets, including a base (31), a bracket (32), a second industrial camera (33), and a second industrial lens (34). The base (31) is fixedly installed on the upper surface edge of the upper frame near the air damper. The base (31) is a telescopic base. The bracket (32) is located on the upper surface of the base (31). The second industrial camera (33) is fixedly connected to the bracket (32). The second industrial lens (34) is connected to the second industrial camera (33). The number of the second image recognition components (3) is the same as the number of the air damper of the hydro turbine generator. The second image recognition component (3) is installed at the location of each air damper.

2. The image recognition-based hydro-generator fault accurate identification system according to claim 1, characterized in that, The outer side of the annular track (11) is designed with a groove, and gear columns (111) are evenly distributed in the groove on the outer side of the annular track (11). The motion support (12) includes a sliding plate (121), a motor (122), a gear (123), a camera mounting bracket (124), and a pulley (125). The sliding plate (121) is slidably connected to the upper surface of the annular track (11). A groove is provided on one side above the sliding plate (121), and the motor (122) is installed in the groove. The output shaft of the motor (122) passes through the sliding plate (121) and connects to the groove. The gear (123) is adapted to the gear post (111). The camera mounting bracket (124) is fixedly connected to the upper surface of the sliding plate (121) and located to the right of the motor (122). There are four pulleys (125), which are fixedly connected to the front and rear sides of the sliding plate (121). The two pulleys (125) on the side of the groove of the annular track (11) of the sliding plate (121) are located on the left and right sides of the gear (123), respectively. The two pulleys (125) on the other side are tangent to the inner surface of the annular track (11).

3. The image recognition-based hydro-generator fault accurate identification system according to claim 1, characterized in that, The second image recognition component (3) also includes a visual light source (35) and a light source controller (36). The bracket (32) has an S-shaped structure. The second industrial camera (33) is fixedly installed on the upper surface of the bracket (32). The visual light source (35) is installed below the second industrial camera (33). The light source controller (36) is connected to the visual light source (35) and is located on the right side of the visual light source (35). The visual light source (35) is a strip light source.

4. The method for determining faults in a hydro-generator based on image recognition as described in any one of claims 1-3, characterized in that, Includes the following steps: S01: Data acquisition, acquiring the status of the shear pin in the shear pin signal device (2), the status of the wind gate sensor (4), the image of the first image recognition component (1), and the image of the second image recognition component (3); S02: Data processing, The angle between the guide vane crank arm and the connecting rod is calculated in the image processing component (5) based on the image captured by the first image recognition component (1). , The distance between the upper surface of the wind gate and the lower surface of the turbine generator brake ring is calculated in the image processing component (5) using the image captured by the second image recognition component (3). ; S03: Fault diagnosis, the diagnosis process is as follows: When detecting image features When the angle is less than or equal to the preset angle threshold, it is determined that the feature is consistent with the feature in the preset image feature library of the normal position state of the shear pin. When detecting image features When the angle exceeds the preset threshold, it is determined that the feature comparison with the preset image feature library of the normal position state of the shear pin is inconsistent. When detecting image features When the distance is less than or equal to the preset lifting distance threshold, it is determined that the feature is consistent with the feature in the preset windshield gate lifting position state image feature library; When detecting image features If the distance exceeds the preset threshold, it is determined that the feature comparison with the preset windshield gate position status image feature library is inconsistent. When detecting image features When the distance is less than or equal to the preset falling distance threshold, it is determined that the feature is consistent with the feature in the preset air damper gate falling position state image feature library; When detecting image features If the distance exceeds the preset threshold, it is determined that the feature is inconsistent with the feature in the preset image feature library of the windshield gate's falling position. S04: Output the results on the display screen (6). The output process is as follows: If the cut pin is in the cut state, the image features... If the feature does not match the preset image feature library, the output will be: Cut off by the X-shaped cutting pin. If the cut pin is in an uncut state, the image features... If the feature does not match the feature in the preset image feature library, the output will be: "X-shaped cutting pin may have been cut, but no signal is reported at this time." If the cut pin is in the cut state, the image features... When the feature matches the feature in the preset image feature library, output: false alarm of cutting signal of cutting pin X; If the cut pin is in an uncut state, the image features... When the feature matches the feature in the preset image feature library, the output is: "X-numbered shear pin is in normal condition." If the state of the windshield sensor (4) is raised, that is, the upper sensor (41) detects a pressure signal, and the position of the windshield plate matches the preset windshield plate raised position state image feature library, the output is: Windshield X is in normal raised state. If the state of the wind gate sensor (4) is raised, that is, the upper sensor (41) detects a pressure signal, and the position of the wind gate plate is inconsistent with the preset wind gate plate raised position state image feature library, the output is: Wind gate X is in an abnormal state. If the state of the windshield sensor (4) is falling, that is, the lower sensor (42) detects a pressure signal, and the position of the windshield plate matches the preset windshield plate falling position state image feature library, the output is: the falling state of windshield X is normal. If the state of the wind gate sensor (4) is falling, that is, the lower sensor (42) detects a pressure signal, and the position of the wind gate plate is inconsistent with the preset wind gate plate falling position state image feature library, the output is: Wind gate X is in an abnormal state. If the state of the windshield sensor (4) is not raised or lowered, that is, neither the upper sensor (41) nor the lower sensor (42) detects a pressure signal, and the position of the windshield plate is inconsistent with the preset windshield plate raised or lowered position state image feature library, the output is: the position of windshield X is abnormal. If the state of the windshield sensor (4) is not raised or lowered, that is, neither the upper sensor (41) nor the lower sensor (42) detects a pressure signal, and the position of the windshield plate is inconsistent with the preset windshield plate middle position state image feature library, the output is: Windshield No. X is in an abnormal state. The angle threshold is 17°, the lifting distance threshold is 35mm, and the falling distance threshold is 5mm.

Images