Power station video inspection strategy generation method based on digital twinning

Abstract

The invention provides a power station video inspection strategy generation method based on digital twinning, and the method comprises the following steps: S1, supplementing an inspection point and an industrial television system camera model based on a three-dimensional model, and completing a power station digital twinning model; s2, carrying out attribute marking on each inspection point and a camera of the industrial television system; s3, carrying out spatial visibility analysis of the industrial television camera on the inspection point, and recording the inspection point meeting spatial visibility into an inspection point-camera two-tuple; and S4, generating a camera preset point of the industrial television system, and forming an inspection point-camera-preset point triple inspection strategy. According to the invention, on the basis of the spatial analysis capability of the three-dimensional model, the automatic analysis of many-to-many spatial visibility between the industrial television system camera and the inspection point, the automatic calculation of the camera preset point and the automatic generation of the inspection point-camera-preset point triple inspection strategy are realized; the inspection strategy generation and strategy configuration efficiency is improved, and defects existing in manual configuration can be avoided.

Description

Technical Field

[0001] This invention belongs to the field of equipment inspection technology, specifically relating to a method for generating power plant video inspection strategies based on digital twins. Background Art

[0002] As a key energy storage facility in the new power system, pumped storage power stations are crucial for ensuring the absorption of new energy sources and the stable operation of the power system. Regular inspections are an effective means for power stations to promptly grasp the operating status of equipment and ensure its safe and stable operation; therefore, they are routinely carried out in pumped storage power stations. Currently, inspections are generally conducted by staff carrying paper inspection forms or handheld inspection terminals, following a pre-set inspection route to manually inspect and record the equipment that needs to be inspected.

[0003] With the construction of new power systems, pumped storage hydropower has entered a stage of large-scale development, leading to a contradiction between the high demand for personnel and the shortage of professional staff in operation and maintenance. Therefore, gradually reducing on-site personnel through digital technology is a common need in the industry. To this end, the industry is currently conducting extensive research on technologies to replace manual inspections, including robotic inspection, drone inspection, and video inspection. Among these, video inspection, which can utilize existing industrial television systems in power plants, will be an essential component of future unmanned inspections. Currently, the recognition algorithms related to video inspection are highly mature, demonstrating high accuracy in recognizing common cabinet and device indicator lights, pressure plate status, handle positions, digital and pointer meter readings (including level gauges), valve positions, switch body open / close positions, oil levels, disconnectors, and grounding switch mechanical position indicators, effectively replacing manual on-site inspections. However, current video inspections mainly employ a method of manually configuring preset camera positions, forming a binary inspection strategy of preset position-inspection point. This approach requires maintenance personnel to manually formulate inspection strategies before conducting video inspections and input the corresponding preset positions into the industrial television system. These preset positions include attributes such as camera pose and zoom level, with pose including horizontal and tilt angles. Therefore, maintenance personnel need to manually assess and test preset attributes for hundreds of industrial television cameras and hundreds or even thousands of inspection points on-site, identifying and establishing many-to-many relationships between cameras and inspection points. This can result in thousands of preset positions being configured, placing a heavy workload on on-site maintenance personnel and making it highly susceptible to omissions, misconfigurations, incorrect recordings, and other errors.

[0004] Chinese Patent Publication No. CN118233597A discloses a method, system, and storage medium for intelligent inspection of hydropower stations. This invention utilizes intelligent video inspection of key equipment and areas, capturing and identifying equipment in key areas according to the inspection process to filter out hazardous areas. When inspection personnel enter a hazardous area, timely boundary alarms or inspection precaution reminders are provided, significantly improving the efficiency and intelligence of inspection tasks. While this invention proposes setting preset camera positions according to actual scenario needs, it does not address how to set preset positions for industrial television system cameras. Chinese Patent Publication No. CN111597231A discloses a substation intelligent inspection system based on multi-source heterogeneous system data mining. This system achieves linked inspection based on equipment IDs in the production management system by integrating multiple systems. However, it also does not address how to set preset points for industrial television cameras and generate inspection strategies, and requires manual definition of the preset points to be inspected. It is evident that while existing technologies have proposed video inspection systems and methods, they have not yet proposed a method for generating video inspection strategies for industrial television systems. Manual configuration will introduce a large amount of workload and will inevitably lead to omissions, misconfigurations, incorrect recordings, and other issues, which will directly affect the efficiency, quality, and even feasibility of video inspection. Summary of the Invention

[0005] The main objective of this invention is to provide a method for generating power plant video inspection strategies based on digital twins, addressing the aforementioned problems.

[0006] Therefore, the above-mentioned objective of the present invention is achieved through the following technical solution: A method for generating a power plant video inspection strategy based on digital twins includes the following steps: S1. Based on the completed 3D model of the power plant, supplement the inspection points and the camera models of the industrial television system to improve the digital twin model of the power plant. S2. Mark the attributes of each inspection point and industrial television system camera within the digital twin model of the power plant; S3. Based on the digital twin model of the power plant, conduct spatial visibility analysis of industrial TV cameras on inspection points, including visual angle visibility and field of view visibility, and record the spatial visibility in the inspection point-camera binary. S4. Based on the spatial visibility analysis results, generate preset points for industrial television system cameras and form a three-element inspection strategy of inspection point-camera-preset point.

[0007] While adopting the above technical solutions, the present invention may also adopt or combine the following technical solutions: As a preferred embodiment of the present invention: in step S2, the inspection point attributes include at least the equipment code, the type of inspection object, the spatial plane, the plane size, and the viewing angle.α。

[0008] As a preferred technical solution of the present invention: the inspection point attribute also includes a normal vector, the starting point of the normal vector is the center point of the spatial plane, and the normal vector is perpendicular to the spatial plane and points inward.

[0009] As a preferred embodiment of the present invention: in step S2, the camera attributes include at least device code, initial position, and sensor size. h Aperture value N Zoom range, gimbal horizontal rotation angle range, gimbal pitch rotation angle range, viewing angle range, and viewing distance range.

[0010] As a preferred technical solution of the present invention: the viewing angle range of the camera includes a horizontal viewing angle range and a vertical viewing angle range. The horizontal viewing angle range is determined by combining the initial normal vector and the horizontal rotation angle range of the gimbal. The vertical viewing angle range is determined by combining the initial normal vector and the pitch rotation angle range. The initial normal vector is obtained based on the initial position of the camera.

[0011] As a preferred embodiment of the present invention: the camera's viewing distance range is the closest viewing distance. L min and maximum visible distance L max The range between them is calculated using the following formula: L min = uf min 2 / ( f min 2 + uNC ); L max = uf max 2 / ( f max 2 - uNC ); C = h / 1000; In the formula, u For object distance, f min For the minimum focal length, f min For maximum focal length, N This is the aperture value. C To allow for the radius of the dispersion circle, h This refers to the sensor size.

[0012] As a preferred embodiment of the present invention, step S3 further includes the following sub-steps: S31. Determine whether the inspection point is within the camera's field of view. If not, the field of view is not visible. Move to the next inspection point and repeat step S31. Otherwise, proceed to the next step. S32. Construct a spatial vector between the inspection point and the industrial television system camera. The direction of the spatial vector is the center point of the spatial plane of the inspection point pointed by the industrial television system camera. Determine whether the spatial vector is within the viewing angle range of the industrial television system camera. If not, the viewing angle is not visible. Proceed to the next inspection point and return to step S31. Otherwise, proceed to the next step. S33. Calculate the angle α between the normal vector of the inspection point and the spatial vector. If 180°-a>α, the viewpoint is not visible. Proceed to the next inspection point and return to step S31. Otherwise, proceed to the next step. S34. Perform collision analysis on spatial vectors within the digital twin model of the power station. If a collision point exists, the field of view is not visible. Proceed to the next inspection point and return to step S31. Otherwise, record it in the inspection point-camera binary pair, proceed to the next industrial television system camera, and return to step S31 until the spatial visibility analysis of all inspection points and cameras is completed.

[0013] As a preferred embodiment of the present invention, step S4 further includes the following sub-steps: S41. For the camera-inspection point pair, construct the spatial vector of the camera and the inspection point. The direction of the spatial vector is the point from the camera to the center of the spatial plane of the inspection point. S42. Calculate the horizontal rotation angle and the pitch rotation angle of the gimbal based on the spatial vector. S43. Calculate the focal length based on the distance between the camera and the inspection point. f ,like f Greater than the maximum focal length f max or less than the minimum focal length f min Then proceed to the next camera-inspection point pair and return to step S41, as shown in the following formula: f =2 uH / h ; In the formula, u The distance between the camera and the inspection point. H The diagonal length of the inspection point's spatial plane. h For sensor dimensions; S44. Configure the gimbal horizontal rotation angle, gimbal pitch rotation angle and focal length as preset points to form a three-element inspection strategy of inspection point-camera-preset point. Proceed to the next two-element camera-inspection point and return to step S41.

[0014] As a preferred technical solution of the present invention, it further includes step S5, optimization of the inspection strategy for inspection points, which specifically includes the following sub-steps: S51. For each inspection point-camera-preset point triplet inspection strategy belonging to an inspection point, calculate the spatial perspective. b The formula is as follows: b =180°- a ; In the formula, a The angle between the spatial vector between the inspection point and the camera and the normal vector of the inspection point is given by the direction of the spatial vector, which points from the camera to the center of the spatial plane of the inspection point. S52. Define the inspection strategy with the smallest spatial perspective as the first inspection strategy for that inspection point. S53. If the distance between the camera and the inspection point in the first inspection strategy is greater than the median value of the camera's visible distance range and there are other inspection strategies, then the inspection strategy with the second smallest spatial perspective is defined as the second inspection strategy for that inspection point.

[0015] Compared with the prior art, the present invention has the following beneficial effects: Based on the spatial analysis capabilities of 3D models, this invention enables automatic analysis of many-to-many spatial visibility between cameras and inspection points in industrial television systems, as well as automatic calculation of camera preset points and automatic generation of inspection point-camera-preset point triplet inspection strategies. This improves the efficiency of inspection strategy generation and configuration, and avoids situations such as omissions, mismatches, omissions, and misrecordings that occur with manual configuration. Attached Figure Description

[0016] Figure 1 The flowchart illustrates the power plant video inspection strategy generation method based on digital twins provided by this invention.

[0017] Figure 2 A flowchart illustrating the steps for optimizing the inspection strategy by increasing the number of inspection points. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0019] like Figure 1 As shown, a method for generating a power plant video inspection strategy based on digital twins specifically includes the following steps: S1. Based on the completed 3D model of the power plant, supplement the inspection points and the camera models of the industrial television system to improve the digital twin model of the power plant. S2. Mark the attributes of each inspection point and industrial television system camera within the digital twin model of the power plant; The attributes of an inspection point should include at least the equipment code, the type of inspection object, the spatial plane, the plane dimensions, and the viewing angle. α。

[0020] Taking a power cabinet inspection point in the chiller room of a pumped storage power station as an example, the types of objects to be inspected at this inspection point include status indicator lights, rectangular digital displays, square pointer meters, and on / off switches. The visibility angle of the indicator lights is as high as 70 degrees, the visibility angle of the digital displays is as high as 50 degrees, the visibility angle of the pointer meters is only 5 degrees, and the visibility angle of the on / off switches is also as high as 70 degrees. Therefore, the visibility angle of this inspection point is 5 degrees.

[0021] The inspection point attributes also include the normal vector. The starting point of the normal vector is the center point of the spatial plane. The normal vector is perpendicular to the spatial plane and points inward. The spatial plane is the front surface of the power cabinet.

[0022] The visual angle of the inspection point is determined according to the type of the inspection object, and the minimum value of the visual angle of each inspection object within the inspection point is taken as the visual angle of the inspection point.

[0023] Camera attributes should include at least device code, initial position, and sensor size. h Aperture value N Zoom range, gimbal horizontal rotation angle range, gimbal pitch rotation angle range, viewing angle range, and viewing distance range.

[0024] The camera's viewing angle range includes the horizontal viewing angle range and the vertical viewing angle range. The horizontal viewing angle range is determined by combining the initial normal vector and the pan-tilt unit's horizontal rotation angle range. The vertical viewing angle range is determined by combining the initial normal vector and the pitch rotation angle range. The initial normal vector is obtained based on the camera's initial position, which is generally provided by the installation unit, including the installation height and installation angle. If the installation unit does not provide this information, it can be obtained by measuring on-site.

[0025] Currently, commonly used cameras in power plants mainly include two types: bullet cameras and PTZ cameras. Bullet cameras only offer zoom capability and do not support rotation; therefore, the horizontal and vertical rotation angles of the PTZ unit are both 0. The sensor size... h Aperture value N The zoom range and zoom range are parameters of the camera lens.

[0026] The camera's field of view range is the closest visible distance. L min and maximum visible distance L maxThe range between them is calculated using the following formula: L min = uf min 2 / ( f min 2 + uNC ); L max = uf max 2 / ( f max 2 - uNC ); C = h / 1000; In the formula, u For object distance, f min For the minimum focal length, f min For maximum focal length, N This is the aperture value. C To allow for the radius of the dispersion circle, h This refers to the sensor size.

[0027] Currently, the PTZ cameras from mainstream video surveillance manufacturers such as Hikvision and Dahua, which are commonly used in power plants, have a maximum viewing distance of tens or even hundreds of meters.

[0028] S3. Based on the digital twin model of the power plant, conduct spatial visibility analysis of industrial TV cameras on inspection points, including visual angle visibility and field of view visibility, and record the spatial visibility in the inspection point-camera binary. Visuality of the field of view mainly analyzes whether the inspection point is within the camera's field of view and whether the azimuth angle between the camera and the inspection point is less than the inspection point's field of view. Visuality of the field of view mainly analyzes whether the inspection point is within the camera's field of view and whether there are any obstructions between them.

[0029] S31. Determine whether the inspection point is within the camera's line of sight. L min , L max If not (the distance between the two) L < L min or L > L max If the field of view is not visible, proceed to the next inspection point and repeat step S31; otherwise, proceed to the next step. S32. Construct a spatial vector between the inspection point and the industrial television system camera to represent the azimuth angle of the inspection point relative to the camera. The direction of the spatial vector is from the industrial television system camera to the center point of the spatial plane of the inspection point. Determine whether the spatial vector is within the viewing angle range of the industrial television system camera. If the azimuth angle between the two is greater than the viewing angle of the inspection point, it means that the video frame obtained by the camera cannot be used to accurately identify the status or numbers of each inspection object on the inspection. Therefore, the camera is not visible to the inspection point. Move to the next inspection point and return to step S31. Otherwise, proceed to the next step. S33. Calculate the angle α between the normal vector of the inspection point and the spatial vector. If 180°-a>α, then the azimuth angle between the camera and the inspection point is greater than the visible angle of the inspection point. The video frame obtained by the camera cannot be used to accurately identify the status or numbers of each inspection object on the inspection. Therefore, the camera is not visible to the inspection point. Move to the next inspection point and return to step S31. Otherwise, proceed to the next step. S34. Perform collision analysis on spatial vectors within the digital twin model of the power station. If a collision point exists, it indicates that the camera cannot obtain a complete and accurate video frame of the inspection point, and the camera is not visible to the inspection point. Proceed to the next inspection point and return to step S31. Otherwise, record it in the inspection point-camera binary pair, proceed to the next industrial television system camera, and return to step S31 until the spatial visibility analysis of all inspection points and cameras is completed.

[0030] S4. Based on the spatial visibility analysis results, generate preset points for industrial television system cameras and form a three-element inspection strategy of inspection point-camera-preset point.

[0031] S41. For the camera-inspection point pair, construct the spatial vector of the camera and the inspection point. The direction of the spatial vector is the point from the camera to the center of the spatial plane of the inspection point. S42. Calculate the horizontal rotation angle and the pitch rotation angle of the gimbal based on the spatial vector. S43. Calculate the focal length based on the distance between the camera and the inspection point. f ,like f Greater than the maximum focal length f max or less than the minimum focal length f min Then proceed to the next camera-inspection point pair and return to step S41, as shown in the following formula: f =2 uH / h ; In the formula, u The distance between the camera and the inspection point. H The diagonal length of the inspection point's spatial plane. hFor sensor dimensions; S44. Configure the gimbal horizontal rotation angle, gimbal pitch rotation angle and focal length as preset points to form a three-element inspection strategy of inspection point-camera-preset point. Proceed to the next two-element camera-inspection point and return to step S41.

[0032] In actual use, temporary obstructions may occur between the camera and the inspection point, causing the automatically generated inspection strategy to fail to complete the inspection point identification and inspection based on image recognition. Therefore, multiple inspection strategies can be further configured according to the characteristics of the inspection strategy itself, such as... Figure 2 As shown, it also includes step S5, optimization of the inspection strategy for inspection points: S51. For each inspection point-camera-preset point triplet inspection strategy belonging to an inspection point, calculate the spatial perspective. b The formula is as follows: b =180°- a ; In the formula, a The angle between the spatial vector between the inspection point and the camera and the normal vector of the inspection point is given by the direction of the spatial vector, which points from the camera to the center of the spatial plane of the inspection point. S52. Define the inspection strategy with the smallest spatial perspective as the first inspection strategy for that inspection point. S53. If the distance between the camera and the inspection point in the first inspection strategy is greater than the median value of the camera's visible distance range and there are other inspection strategies, then the inspection strategy with the second smallest spatial perspective is defined as the second inspection strategy for that inspection point.

[0033] After the inspection strategies for each inspection point are generated, the power plant can call each inspection strategy to complete the inspection work of its respective inspection point after the actual inspection task is started. If the image recognition model in the background of the inspection system finds that it cannot accurately identify the status or value of the inspection object at a certain inspection point, the second inspection strategy of that inspection point can be called to supplement the inspection, thereby completing the inspection work of each inspection point.

[0034] The technical solution of the present invention has been described in conjunction with the specific experimental procedures shown in the accompanying drawings. However, the scope of protection of the present invention is not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions resulting from such changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for generating power plant video inspection strategies based on digital twins, characterized in that, Includes the following steps: S1. Based on the completed 3D model of the power plant, supplement the inspection points and the camera models of the industrial television system to improve the digital twin model of the power plant. S2. Mark the attributes of each inspection point and industrial television system camera within the digital twin model of the power plant; S3. Based on the digital twin model of the power plant, conduct spatial visibility analysis of industrial TV cameras on inspection points, including visual angle visibility and field of view visibility, and record the spatial visibility in the inspection point-camera binary. S4. Based on the spatial visibility analysis results, generate preset points for industrial television system cameras and form a three-element inspection strategy of inspection point-camera-preset point.

2. The method according to claim 1, characterized in that: In step S2, the inspection point attributes include at least the equipment code, the type of inspection object, the spatial plane, the plane size, and the viewing angle. α。 3. The method according to claim 1 or 2, characterized in that: The inspection point attributes also include the normal vector, whose starting point is the center point of the spatial plane, and the normal vector is perpendicular to the spatial plane and points inward.

4. The method according to claim 1, characterized in that: In step S2, the camera attributes include at least the device code, initial position, and sensor size. h Aperture value N Zoom range, gimbal horizontal rotation angle range, gimbal pitch rotation angle range, viewing angle range, and viewing distance range.

5. The method according to claim 4, characterized in that: The camera's viewing angle range includes the horizontal viewing angle range and the vertical viewing angle range. The horizontal viewing angle range is determined by combining the initial normal vector and the pan-tilt unit's horizontal rotation angle range. The vertical viewing angle range is determined by combining the initial normal vector and the pitch rotation angle range. The initial normal vector is obtained based on the camera's initial position.

6. The method according to claim 4, characterized in that: The camera's field of view range is the closest visible distance. L min and maximum visible distance L max The range between them is calculated using the following formula: L min = uf min 2 / ( f min 2 + uNC )； L max = uf max 2 / ( f max 2 - uNC )； C = h / 1000； In the formula, u For object distance, f min For the minimum focal length, f min For maximum focal length, N This is the aperture value. C To allow for the radius of the dispersion circle, h This refers to the sensor size.

7. The method according to claim 1, characterized in that: Step S3 also includes the following sub-steps: S31. Determine whether the inspection point is within the camera's field of view. If not, the field of view is not visible. Move to the next inspection point and repeat step S31. Otherwise, proceed to the next step. S32. Construct a spatial vector between the inspection point and the industrial television system camera. The direction of the spatial vector is the center point of the spatial plane of the inspection point pointed by the industrial television system camera. Determine whether the spatial vector is within the viewing angle range of the industrial television system camera. If not, the viewing angle is not visible. Proceed to the next inspection point and return to step S31. Otherwise, proceed to the next step. S33. Calculate the angle α between the normal vector of the inspection point and the spatial vector. If 180°-a>α, the viewpoint is not visible. Proceed to the next inspection point and return to step S31. Otherwise, proceed to the next step. S34. Perform collision analysis on spatial vectors within the digital twin model of the power station. If a collision point exists, the field of view is not visible. Proceed to the next inspection point and return to step S31. Otherwise, record it in the inspection point-camera binary pair, proceed to the next industrial television system camera, and return to step S31 until the spatial visibility analysis of all inspection points and cameras is completed.

8. The method according to claim 1, characterized in that: Step S4 also includes the following sub-steps: S41. For the camera-inspection point pair, construct the spatial vector of the camera and the inspection point. The direction of the spatial vector is the point from the camera to the center of the spatial plane of the inspection point. S42. Calculate the horizontal rotation angle and the pitch rotation angle of the gimbal based on the spatial vector. S43. Calculate the focal length based on the distance between the camera and the inspection point. f ,like f Greater than the maximum focal length f max or less than the minimum focal length f min Then proceed to the next camera-inspection point pair and return to step S41, as shown in the following formula: f =2 uH / h ； In the formula, u The distance between the camera and the inspection point. H The diagonal length of the inspection point's spatial plane. h For sensor dimensions; S44. Configure the gimbal horizontal rotation angle, gimbal pitch rotation angle and focal length as preset points to form a three-element inspection strategy of inspection point-camera-preset point. Proceed to the next two-element camera-inspection point and return to step S41.

9. The method according to claim 1, characterized in that: It also includes step S5, optimization of the inspection strategy for inspection points, which specifically includes the following sub-steps: S51. For each inspection point-camera-preset point triplet inspection strategy belonging to an inspection point, calculate the spatial perspective. b The formula is as follows: b =180°- a ； In the formula, a The angle between the spatial vector between the inspection point and the camera and the normal vector of the inspection point is given by the direction of the spatial vector, which points from the camera to the center of the spatial plane of the inspection point. S52. Define the inspection strategy with the smallest spatial perspective as the first inspection strategy for that inspection point. S53. If the distance between the camera and the inspection point in the first inspection strategy is greater than the median value of the camera's visible distance range and there are other inspection strategies, then the inspection strategy with the second smallest spatial perspective is defined as the second inspection strategy for that inspection point.

Images