Dam access door detection operation method

Abstract

The invention discloses a dam access door detection operation method, and belongs to the technical field of water conservancy intelligent operation and maintenance. After the quick door is closed, the water flow channel is emptied and the access door is closed, the mooring type spherical unmanned aerial vehicle is lowered into the water flow channel through a dam crest vent hole, and cooperative detection is executed; dry shrinkage cracks and erosion damage are distinguished through texture anisotropy; risk enhancing factors are superimposed on leaks across the metal-rubber interface. And in the passing process, the lowering speed and the cable tension are dynamically adjusted according to the local curvature of the vent hole, and an accumulated curvature self-adaptive segmented lifting strategy is adopted in the recycling process. According to the invention, the problems of bent pore passage passing, leakage misjudgment, unclear maintenance priority and the like are solved, and safe, accurate and full-automatic detection is realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology for water conservancy and hydropower projects, and in particular to a method for detecting dam inspection gates. Background Technology

[0002] In large-scale water conservancy projects, such as ship locks, floodgates, and hydropower station water diversion systems, inspection gates are crucial structures for ensuring equipment maintenance and emergency sealing. Traditional inspection methods rely on manual entry into the waterway for visual inspection, which suffers from high risk, low efficiency, and numerous blind spots.

[0003] In recent years, drone inspections have been gradually applied, but they face the following problems in water conservancy scenarios: First, most ventilation holes are downward-curving concrete channels, such as J-shaped ones, which are not straight and cannot be traversed by ordinary drones; Second, water channels are enclosed spaces, and wireless signals are severely attenuated, making it difficult for battery-powered drones to guarantee communication and endurance.

[0004] Currently, although tethered drones are used for power line inspection, they operate in open spaces above ground. There is currently no truly feasible solution for inspecting dam inspection gates. Therefore, we propose a method for inspecting dam inspection gates in this application. Summary of the Invention

[0005] The purpose of this invention is to provide a method for inspecting dam maintenance gates, so as to enable the inspection of dam maintenance gates through ventilation holes.

[0006] To achieve the above objectives, this application provides a method for inspecting dam maintenance gates, which employs an integrated inspection system. The integrated inspection system includes an operating platform, on which a spherical drone is connected via a retractable cable. The spherical drone is equipped with a camera and a lidar. The detection method includes the following steps: S1. Close the quick-closing door to drain water from the downstream side of the water channel; S2. Close the inspection door and open the high-speed door; S3. Deploy the integrated detection system on the top of the dam body and position the spherical drone directly above the inlet of the ventilation hole; S4. Based on the pre-acquired 3D model of the vent, the control cable is released with a variable speed strategy, so that the spherical UAV is smoothly lowered into the water channel along the curved inner cavity of the vent. S5. Use a spherical drone equipped with a camera and lidar to conduct collaborative inspection of inspection doors and water channels; S6. After the inspection is completed, the spherical drone is recovered along the original path.

[0007] Once the spherical drone enters the water channel, it performs the following collaborative detection: a. Inspection door leakage detection: The surface of the inspection door is scanned using visible light and polarization imaging to identify water films with mirror reflection, continuous water droplets, or strip-shaped wet spots from top to bottom, and leakage is determined accordingly. b. Inspection of the surface condition of the access door: By fusing laser point cloud data with images, the system identifies metal rust bulges and gaps in the waterproofing rubber. c. Surface inspection of concrete in the flow channel: transversely scan the inner wall of adjacent concrete to identify shrinkage cracks, erosion pits or calcium precipitation.

[0008] In S4, the greater the local curvature of the vent and the larger the radius of the spherical drone, the smaller the safety gap and the higher the risk of jamming. When the risk of jamming exceeds the threshold, the descent speed is automatically reduced and the cable tension is increased to improve the effective passage ability of the spherical drone in the curved channel.

[0009] The maximum permissible lowering speed of the spherical drone within the vent decreases exponentially with the increase of local curvature, ensuring significant speed reduction in sharp bends to maintain stability during passage.

[0010] In S4, the lateral friction force on the cable in the current bend is estimated in real time. This friction force is proportional to the cable tension and the total curvature integral of the bend. If the estimated friction force exceeds the safety threshold, the lowering is paused and the cable is retracted and extended in small steps to release the accumulated stress. The leakage confidence level is calculated by considering the specular reflection intensity of the wet patch area, the proportion of wetted area, and whether it is distributed in a strip shape along the direction of gravity. If the confidence level reaches the preset high confidence threshold, it is determined to be a real leakage.

[0011] For leakage areas spanning the metal-rubber interface, the risk priority score is further enhanced by an enhancement factor proportional to the intersection length of the defect and the interface, on top of the basic leakage confidence level, to highlight the severity of leakage due to sealing failure.

[0012] When inspecting concrete in water channels, the degree of texture anisotropy is calculated based on the second derivative features of the damaged area image. If the degree of anisotropy is low and the distribution is mesh-like, it is determined to be drying shrinkage cracks; if the degree of anisotropy is high and the distribution is radial, it is determined to be erosion damage.

[0013] If the rusted area on the surface of the inspection door exceeds the rusted area threshold dynamically adjusted based on its service life, or if the density of adjacent concrete cracks exceeds the corresponding service life-related threshold, then targeted maintenance suggestions will be generated respectively.

[0014] In S6, a segmented lifting strategy is adopted in the recycling phase. The pause time of each segment is dynamically extended according to the cumulative curvature of that segment and all previous curved segments, so that a longer cable relaxation time is given in the high curvature accumulation area, so that the flexible cable can naturally conform to the hole wall of the vent under the action of gravity.

[0015] Compared with the prior art, the above-conceptual technical solution conceived in this application has the following beneficial effects: 1. This application provides a method for inspecting dam maintenance gates, eliminating the high-risk operational scenarios associated with traditional manual entry into water channels for inspection. Operators no longer need to enter narrow, damp, and enclosed water channels, avoiding major safety hazards such as falls from heights, oxygen deficiency, and sudden water ingress. The entire process of drilling, inspection, and retrieval is completed remotely via a spherical drone.

[0016] 2. This invention employs a multi-feature fusion discrimination mechanism based on visible light and polarization imaging. It integrates the intensity of specular reflection, the proportion of wetted area, and whether it exhibits a top-down gravity stripe morphology to construct a Leakage Confidence Index (LCI). Only when the index exceeds a high confidence threshold is it determined to be a real leak, which significantly suppresses false alarms caused by condensate or oil. For leaks that cross the interface between the metal door leaf and the rubber waterstop, the system automatically adds a risk enhancement factor to accurately identify the most dangerous sealing failure defects, ensuring that no critical hidden dangers are overlooked.

[0017] 3. This invention establishes a curvature-tension coupling passage control method and utilizes the mechanical mechanism that increasing cable tension can effectively improve the passage clearance of a spherical UAV in curves. Combined with exponential decay descent speed control and online friction estimation and jitter escape strategy, this invention enables a flexible tethered spherical UAV to stably and reliably pass through J-shaped or S-shaped concrete ventilation holes, which is impossible for ordinary rotary-wing UAVs.

[0018] 4. This invention analyzes the second derivative features of the image of the damaged area on the concrete surface and calculates the texture anisotropy index (TAI), effectively distinguishing between harmless shrinkage cracks and harmful high-speed water erosion pits. Combined with laser point cloud data, it can also identify conditions such as metal corrosion bulges and detachment of the water-stop rubber. Based on the service life of the gate, the evaluation threshold is dynamically adjusted to automatically generate targeted maintenance suggestions such as "replace the water-stop rubber" and "repair the erosion area", which greatly improves the scientific nature of operation and maintenance decisions.

[0019] 5. The present invention adopts a segmented lifting mechanism based on curvature accumulation and adaptation. The pause time of each segment is dynamically extended with the total deflection angle of the bend that has been passed, so that the flexible cable naturally fits the hole wall under the action of gravity, fully releases stress, and effectively prevents jamming or snagging during the recycling process. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0021] Figure 1This is a schematic diagram of the integrated detection system of the present invention.

[0022] Figure 2 This is a diagram showing the state of the fast door when it is closed, as used in the application of this invention.

[0023] Figure 3 This is a diagram showing the state of the maintenance door being closed and the high-speed door being opened when the present invention is applied.

[0024] Figure 4 This is a diagram showing the state of a spherical drone entering a water channel through a vent when the present invention is applied.

[0025] Figure label: 1. Dam body; 2. Dam crest; 3. Water flow channel; 4. Quick-access door; 5. Inspection door; 6. Ventilation hole. Operating platform 10, cables 20, spherical drone 30. Detailed Implementation

[0026] To more clearly illustrate the purpose, technical solution, and beneficial effects of this application, a further detailed description of this application is provided below in conjunction with illustrations and specific embodiments. It should be specifically noted that the specific embodiments described below are only for illustrating the technical content of this application and do not constitute a limitation on the scope of protection of this application.

[0027] Regarding the explanation of terminology: In this application, "and / or" is used to describe the relationship between related objects, covering three possible situations: taking "A and / or B" as an example, it can indicate the situation where only A exists, A and B exist simultaneously, or only B exists; the symbol " / " indicates the "or" relationship between related objects, such as "A / B" which refers to A or B.

[0028] Regarding the description of the embodiments: The terms "exemplary" and "for example" appearing in this application are only used to illustrate the technical solutions through specific examples. It should be particularly emphasized that any implementation method or design scheme marked as "exemplary" or "for example" should not be construed as having an advantage over other solutions. Such expressions are only used to present the technical concepts more intuitively.

[0029] Example 1: See Figure 2 The dam body 1 is equipped with an inspection gate 5 on the upstream side and a quick gate 4 on the downstream side. The quick gate 4 and the inspection gate 5 are used to open and close the water flow channel 3. The dam body 1 is equipped with a vent 6 on the downstream side of the quick gate 4, which connects the dam top 2 and the water flow channel 3. The vent 6 is a non-straight curved channel that extends downward along the direction of gravity.

[0030] See Figure 1This invention provides a method for inspecting dam maintenance gates, which employs an integrated inspection system. The integrated inspection system includes an operating platform 10, on which a spherical drone 30 is connected via a retractable cable 20. The spherical drone 30 is equipped with a camera and a lidar. The operating platform 10 is used to control the telescopic arm, the retraction and extension of the cable 20, and the power supply and control of the spherical drone 30.

[0031] The detection method includes the following steps: S1. Close the quick-closing door 4, and the water downstream of the water channel 3 will be discharged; S2. Close the maintenance door 5 and open the high-speed door 4; S3. Deploy the integrated detection system on the top 2 of the dam body 1, and position the spherical drone 30 directly above the inlet of the ventilation hole 6; S4. Based on the pre-acquired three-dimensional model of the vent 6, the control cable 20 is released with a variable speed strategy, so that the spherical UAV 30 is smoothly lowered into the water channel 3 along the curved inner cavity of the vent 6. S5. A spherical drone 30 equipped with a camera and lidar is used to conduct collaborative inspection of the inspection door 5 and the water channel 3. S6. After the inspection is completed, recover the spherical drone 30 along the original path.

[0032] The above method eliminates the high-risk operational scenarios associated with traditional manual entry into water channels for inspection. Operators no longer need to enter the narrow, damp, and enclosed water channel 3, avoiding major safety hazards such as falls from heights, lack of oxygen, and sudden water ingress. The entire process of drilling, inspection, and retrieval is completed remotely via a spherical drone.

[0033] Specifically, after the spherical drone 30 enters the water flow channel 3, it performs the following collaborative detection: a. Inspection door leakage detection: The surface of the inspection door 5 is scanned using visible light and polarization imaging to identify water films with mirror reflection, continuous water droplets, or strip-shaped wet spots from top to bottom, and leakage is determined. b. Inspection of the surface condition of the access door: By fusing laser point cloud data with images, the system identifies metal rust bulges and gaps in the waterproofing rubber. c. Surface inspection of concrete in the flow channel: transversely scan the inner wall of adjacent concrete to identify shrinkage cracks, erosion pits or calcium precipitation.

[0034] The above methods together constitute the collaborative detection of the inspection door and its adjacent flow channel structure in this invention. Its function is not only to discover defects, but also to accurately identify the type of defect, locate the root cause of failure, and provide an actionable basis for subsequent maintenance.

[0035] By using visible light and polarization imaging technology to identify reflective water films, continuous water droplets, or strip-shaped wet spots from top to bottom, it can not only effectively distinguish between real leaks and interference items such as condensation and oil stains, but also establish a high-confidence leak criterion based on physical morphology and optical properties, solving the technical problem of difficulty in detecting weak leaks in a waterless drainage environment.

[0036] By fusing laser point cloud data with visible light images to assess the condition of inspection door surfaces, structural defects such as metal rust bulges and gaps in the waterproofing rubber can be accurately identified, revealing the potential root causes of leaks. For example, even if no obvious water marks are currently visible, the presence of a gap between the waterproofing rubber and the pressure plate indicates that the sealing system has failed and there is a risk of leakage, thus shifting from post-incident detection to pre-incident warning.

[0037] By performing a transverse scan of the inner wall of the concrete of adjacent flow channels to identify features such as shrinkage cracks, erosion pits, or calcium precipitation, it is possible not only to distinguish between harmless material shrinkage cracks and harmful high-speed water flow erosion damage, but also to infer historical leakage behavior through chemical traces such as calcium precipitation. This allows for the linking of leakage problems of inspection doors with secondary damage to the flow channel structure, forming a complete chain of evidence for leakage, erosion, and deterioration.

[0038] The synergistic effect of these three detection methods enables maintenance personnel not only to accurately determine whether there is a leak, but also to gain a deeper understanding of why the leak occurred, how long it lasted, and what consequences it caused. This allows them to generate scientific, accurate, and actionable maintenance recommendations, thereby improving the intelligence level of water conservancy facility detection and the efficiency of maintenance decision-making.

[0039] In S4, the larger the local curvature of the vent 6 and the larger the radius of the spherical drone 30, the smaller the safety gap and the higher the risk of jamming. When the risk of jamming exceeds the threshold, the lowering speed is automatically reduced and the tension of the cable 20 is increased to improve the effective passage ability of the spherical drone 30 in the curved channel.

[0040] In the hydraulic monitoring scenario described in this invention, the vent 6 is a downward-curving concrete channel, and its inner wall constitutes a spatial constraint on the spherical UAV 30. The greater the local curvature of the vent 6, i.e., the larger the bend, or the larger the radius of the spherical UAV 30, the smaller the effective passage space available to the spherical UAV 30 on the cross-section of the bend. This is because, in the curved section, the minimum distance between the outer edge of the spherical UAV 30 and the vent wall, i.e., the safety clearance, decreases with the increase of curvature and the size of the spherical UAV 30. Once this clearance is less than a preset safety value, the spherical UAV 30 is highly susceptible to friction with the vent wall or even jamming. To address this risk, when the jamming risk exceeds a threshold, the lowering speed is automatically reduced to decrease inertial disturbance, while the tension of the cable 20 is increased. The increased tension tightens the flexible tether line and makes it fit more closely to the three-dimensional centerline of the vent, thereby pulling the spherical UAV tangentially towards the inside of the bend and obtaining greater lateral margin in the cross-section. This mechanical effect is equivalent to expanding the available passage space, which improves the effective passage capability of spherical drones in curved passages.

[0041] In this embodiment, the expression for the jamming risk criterion is: ; In the formula: This represents the local curvature of the vent at path position s, obtained from a BIM model or laser scanning. Indicates the radius of a spherical unmanned aerial vehicle; Indicates the preset basic safety clearance; Indicates the current cable tension; This represents the tension-stiffness compensation coefficient, which characterizes the reinforcing effect of tension on the equivalent travel radius. The experimental calibration value is generally 0.05 to 0.08 m / N.

[0042] The maximum permissible descent speed of the spherical drone 30 within the ventilation hole 6 decreases exponentially with increasing local curvature, ensuring significant speed reduction in sharp bends to maintain stability. During the descent of the spherical drone 30 along the curved ventilation hole, maintaining high speed in sharp bends, i.e., areas with significant local curvature, would lead to increased inertial force, trajectory deviation from the centerline, increased risk of collision with the hole wall, and even jamming or equipment damage. To ensure stability, the maximum permissible descent speed of the spherical drone 30 should dynamically decrease with increasing local curvature of the ventilation hole. Specifically, an exponential decay law is used for speed control, maintaining higher efficiency in gentle bends or straight sections, while significantly reducing speed in sharp bends, giving the spherical drone 30 sufficient time to respond to guiding forces, such as cable tension, and smoothly conform to the inner wall of the bend. This nonlinear speed control strategy ensures both overall operational efficiency and safety and trajectory accuracy in high-curvature areas.

[0043] In this embodiment, the expression for the control logic is: ; In the formula: This represents the maximum permissible drop speed of the spherical UAV at path position s; This represents the base descent speed, i.e., the ideal speed without curvature; This represents the local curvature of the vent at position s; This represents the curvature suppression coefficient, reflecting the sensitivity of velocity to curvature, with a typical value range of 1.5–2.5 m. This represents a natural exponential function, ensuring that the velocity decays smoothly and rapidly with curvature.

[0044] In S4, the lateral friction force on the cable 20 in the current bending section is estimated in real time. This friction force is proportional to the cable tension and the total curvature integral value of the bending section. If the estimated friction force exceeds the safety threshold, the lowering is paused and the cable 20 is retracted and extended in small steps to release the accumulated stress. During the descent of the spherical drone 30 along the curved vent 6, the cable 20 slides tightly against the wall of the vent 6. Since the vent 6 is a non-linear channel, the cable 20 is subjected to a normal constraint force from the vent wall at the bend, resulting in lateral friction. This friction not only hinders the descent but also accumulates elastic stress within the cable; if it continues to increase, it may cause the cable to become stuck at the bend, leading to jamming or even breakage. Experiments have shown that the magnitude of this lateral friction is proportional to the current cable tension and also proportional to the total deflection angle of the curved section, i.e., the integral value of the curvature along the path. Based on this principle, the system can estimate the friction in real time. Once it exceeds a safety threshold, the descent is immediately paused, and the retraction device is controlled to perform small reciprocating movements of the cable, such as ±0.2 m, to release the accumulated mechanical stress through slight vibrations, restoring the cable to a compliant state and effectively preventing jamming.

[0045] In this embodiment, the expression for the friction estimation model is: ; In the formula: This represents the total lateral frictional force experienced by the cable at the current bend. This represents the coefficient of dynamic friction between the cable outer sheath and the concrete hole wall; The current cable tension is indicated by a tension sensor built into the cable winding and unwinding device, which is acquired in real time. This represents the local curvature of the vent at path position s; Indicates the point from the start of the current curve segment Go to current location The curvature integral is physically represented by the total deflection angle of that segment.

[0046] During dam inspection, water channel 3 is usually emptied and sealed. Relying solely on visible light images to identify water traces can easily lead to misinterpretations of condensation droplets, oil reflections, metal oxide spots, or construction residue as actual leaks, resulting in numerous false alarms.

[0047] Furthermore, the present invention considers the specular reflection intensity of the wet patch area, the proportion of the wetted area, and whether it is distributed in a strip shape along the direction of gravity to calculate its leakage confidence; if the confidence reaches a preset high confidence threshold, it is determined to be a real leakage.

[0048] True leakage has unique physical and morphological characteristics, which are quantitatively assessed through the fusion of multi-dimensional visual evidence. Specifically, true leakage typically manifests as: first, strong specular reflection caused by the smooth surface of the water film; second, having a certain continuous coverage area, not isolated points; and third, forming a strip-like distribution from top to bottom along the vertical direction under the influence of gravity.

[0049] Based on this, the present invention constructs a Leakage Confidence Index (LCI) and performs a weighted calculation based on the above three features. Only when the LCI exceeds a preset high confidence threshold is it determined to be a true leak. This method effectively eliminates static interference terms and significantly improves the accuracy of the judgment.

[0050] In this embodiment, the mathematical expression of the leakage confidence calculation model is as follows: ; In the formula: The leakage confidence index has a value range of [0, 1]. The normalized value (0-1) of specular reflection intensity is obtained through a polarization imaging module. The real water film appears as a dark area in the depolarized image but as a bright area in the polarized reflection image. The ratio of these two values ​​is obtained after normalization. ; This indicates the proportion of pixels in the wet area to the total number of pixels in the suspected area, reflecting the continuity and coverage of the leakage. The gravity strip indicator function is defined as follows: ; Represents the weighting coefficients, satisfying It is optimized through training with historical samples.

[0051] like ,like If so, it is determined to be a genuine leak.

[0052] Furthermore, in the structure of dam inspection gates, the sealing interface between the metal gate leaf and the rubber waterstop is the core of leak prevention. Once leakage occurs here, it usually indicates that the waterstop system has failed, such as due to rubber aging, insufficient compression, or misalignment. This is a structural hazard that must be addressed first. However, general image detection algorithms can only identify water traces and cannot distinguish between: condensation on the metal plate surface (low risk), micro-seepage in the weld (medium risk), and leakage across the metal-rubber interface (high risk, seal failure). Without differentiation, high-risk defects may be submerged among a large number of common defects, leading to a misallocation of maintenance resources.

[0053] This invention, based on experimental results, shows that in over 90% of cases where leakage marks cross the interface between metal and rubber materials, the sealing system experiences functional failure. Therefore, the system not only calculates the basic leakage confidence index (LCI) but also further analyzes the degree of spatial overlap between the leakage area and the critical functional interface, namely the metal-rubber joint. Consequently, for leakage areas crossing the metal-rubber interface, the risk priority score is further enhanced by a factor proportional to the length of the intersection between the defect and the interface, on top of the basic leakage confidence index, to highlight the severity of leaks due to sealing failure.

[0054] Specifically, the Risk Priority Score (RPS) is based on the base LCI, with an enhancement factor that is proportional to the length of the intersection between the leakage defect and the interface. The longer the intersection, the deeper the leakage penetrates into the core sealing area, and the higher the risk. Through this mechanism, the system automatically marks leaks that indicate sealing failure as high priority, ensuring that maintenance personnel pay attention to the most dangerous potential hazards immediately.

[0055] In this embodiment, the mathematical expression of the risk prioritization scoring model is as follows: ; In the formula: This represents a risk priority score, which is dimensionless; the higher the value, the higher the risk. The basic leakage confidence index (0-1) is calculated by the aforementioned multi-feature fusion model. The length of the intersection between the leakage defect area and the metal-rubber interface is represented by an image segmentation and edge matching algorithm. This indicates the total outline length or principal axis length of the entire leakage defect area. Indicates the proportion of the leakage-covered interface; enhancement factor The value range is [1,2], meaning the risk score can be increased to up to twice the base LCI.

[0056] like ,like If so, it is marked as a high-priority seal failure.

[0057] In the inspection of the inner walls of concrete in water flow channels, common surface damage includes shrinkage cracks and erosion damage. Shrinkage cracks are caused by moisture evaporation during the concrete hardening process; they are typically fine, non-directional, and network-like, representing the material's own shrinkage behavior and generally do not affect structural safety. Erosion damage, on the other hand, is caused by long-term scouring by high-speed water, manifesting as localized pits, aggregate spalling, and radial cracks. These cracks have a clear directionality and concentration, and are considered harmful structural damage requiring timely repair. Therefore, shrinkage cracks and erosion damage need to be classified during inspection.

[0058] Through investigation, this invention has revealed significant differences in the local texture directionality between shrinkage cracks and erosion damage. Shrinkage cracks are formed by random contraction, and their image gradients are uniformly distributed in all directions, exhibiting strong isotropy; while erosion damage is formed by directional water flow impact, and the cracks extend along the energy release direction, exhibiting high directional consistency, i.e., strong anisotropy.

[0059] Therefore, this invention utilizes the second derivative of the image (Hessian matrix) to extract local structural features and calculates the texture anisotropy index (TAI). If the TAI value is low and the cracks are mesh-like, they are identified as shrinkage cracks; if the TAI value is high and the cracks are radial, they are identified as erosion damage. This achieves objective and repeatable classification.

[0060] In this embodiment, the texture anisotropy index (TAI) is expressed as follows: ; In the formula: and These are the two eigenvalues ​​of the Hessian matrix in the neighborhood of a certain pixel in the image of the damaged region. The value range of TAI is [0, 1].

[0061] The Hessian matrix is ​​defined as: ; In the formula: Image intensity function The second-order partial derivatives are calculated using a Gaussian second-order derivative filter.

[0062] TAI ≈ 0, that is This indicates that the texture responds similarly in all directions, exhibiting isotropic properties, such as mesh-like shrinkage cracks; TAI ≈ 1, that is This indicates that the texture responds strongly in a certain main direction and is highly anisotropic, such as radial erosion cracks.

[0063] In this embodiment, if TAI < 0.3 and the crack connectivity is a closed network, it is determined to be a shrinkage crack; if TAI > 0.6 and the crack radiates outward from the center point, it is determined to be erosion damage.

[0064] In the management of water conservancy facilities, the degree of deterioration of inspection gates and flow channel concrete is closely related to their service life. For newly built gates, if rust or cracks appear after less than 5 years of service, it may indicate construction defects or material problems, which requires close attention. For old gates, if they have been in service for more than 10 years, a certain degree of rust is a normal aging phenomenon. If a uniform fixed threshold is used, it will lead to excessive alarms and waste maintenance resources.

[0065] This invention dynamically adjusts the threshold values ​​for condition assessment based on the actual service life of the inspection door or concrete structure. Specifically, for the metal surface of the inspection door 5, a rust area threshold is set that moderately increases with the service life; for the adjacent flow channel concrete, a crack density tolerance threshold is set that moderately increases with the service life. When the measured percentage of rust area or crack density exceeds the corresponding dynamic threshold, the system automatically triggers targeted maintenance suggestions, such as "local rust removal and painting" or "crack grouting and sealing." This mechanism reflects respect for the degradation patterns throughout the entire life cycle of hydraulic engineering structures, making the assessment results more scientific and maintenance decisions more accurate.

[0066] After the spherical UAV completes its inspection, it needs to be retrieved from the water channel 3 along the vent 6 to the dam crest 2 via cable 20. Since the vent 6 is a curved channel, the flexible cable 20 will exhibit a non-uniform fit due to gravity contact with the vent wall during descent. If it is pulled continuously at a uniform speed during retrieval, the cable is prone to jamming at high-curvature bends due to stress accumulation, local stacking, or frictional self-locking, especially in areas with multiple bends connected in series, where the cumulative curvature is large, posing a higher risk.

[0067] Therefore, in S6, a segmented lifting strategy is adopted during the recovery phase. Specifically, the path of vent 6 is divided into several segments, such as every 1 meter, with a pause for a certain period after each segment is lifted. This pause time is not fixed but dynamically extended based on the cumulative curvature of the current segment and all previous curved segments, i.e., the total deflection. In areas with large cumulative curvature, such as continuous sharp bends, a longer pause time is given, allowing the flexible cable sufficient time to relax, slide, and naturally conform to the hole wall contour under gravity, thereby releasing internal bending stress and frictional resistance. This mechanism reduces the probability of recovery jamming and ensures a fully automated closed loop.

[0068] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for inspecting dam maintenance gates, characterized in that: An integrated detection system was adopted, which includes an operating platform (10). A spherical drone (30) is connected to the operating platform (10) via a retractable cable (20). The spherical drone (30) is equipped with a camera and a lidar. The detection method includes the following steps: S1. Close the fast door (4) and the water in the downstream side of the water channel (3) is discharged; S2. Close the maintenance door (5) and open the high-speed door (4); S3. Deploy the integrated detection system on the top (2) of the dam body (1) and position the spherical UAV (30) directly above the inlet of the ventilation hole (6); S4. Based on the pre-acquired three-dimensional model of the vent (6), the control cable (20) is released with a variable speed strategy, so that the spherical UAV (30) is smoothly lowered into the water channel (3) along the curved vent (6) cavity. S5. A spherical drone (30) equipped with a camera and lidar is used to conduct collaborative inspection of the inspection door (5) and the water channel (3); S6. After the detection is completed, recover the spherical drone (30) along the original path.

2. The method for inspecting dam maintenance gates according to claim 1, characterized in that: Once the spherical UAV (30) enters the water channel (3), it performs the following collaborative detection: a. Inspection door leakage detection: The surface of the inspection door (5) is scanned by visible light and polarization imaging to identify water film with mirror reflection, continuous water droplets or strip-shaped wet spots from top to bottom, and leakage is judged. b. Inspection of the surface condition of the access door: By fusing laser point cloud data with images, the system identifies metal rust bulges and gaps in the waterproofing rubber. c. Surface inspection of concrete in the flow channel: transversely scan the inner wall of adjacent concrete to identify shrinkage cracks, erosion pits or calcium precipitation.

3. The method for inspecting dam maintenance gates according to claim 1, characterized in that: In S4, the larger the local curvature of the vent (6) and the larger the radius of the spherical UAV (30), the smaller the safety gap and the higher the risk of jamming. When the risk of jamming exceeds the threshold, the lowering speed is automatically reduced and the cable (20) tension is increased to improve the effective passage capability of the spherical UAV (30) in curved channels.

4. A method for inspecting dam maintenance gates according to claim 1 or 3, characterized in that: The maximum permissible descent speed of the spherical UAV (30) within the vent (6) decreases exponentially with the increase of local curvature, ensuring significant deceleration in sharp bends to maintain stability during passage.

5. The method for inspecting dam maintenance gates according to claim 1, characterized in that: In S4, the lateral friction force on the cable (20) in the current bending segment is estimated in real time. This friction force is proportional to the cable tension and the total curvature integral value of the bending segment. If the estimated friction force exceeds the safety threshold, the lowering is paused and the cable (20) is retracted and extended in small steps to release the accumulated stress.

6. The method for inspecting dam maintenance gates according to claim 2, characterized in that: The leakage confidence level is calculated by considering the specular reflection intensity of the wet patch area, the proportion of wetted area, and whether it is distributed in a strip shape along the direction of gravity. If the confidence level reaches the preset high confidence threshold, it is determined to be a real leakage.

7. The method for inspecting dam maintenance gates according to claim 6, characterized in that: For leakage areas spanning the metal-rubber interface, the risk priority score is further enhanced by an enhancement factor proportional to the intersection length of the defect and the interface, on top of the basic leakage confidence level, to highlight the severity of leakage due to sealing failure.

8. The method for inspecting dam maintenance gates according to claim 2, characterized in that: When inspecting the concrete of the water channel (3), the degree of texture anisotropy is calculated based on the second derivative features of the damaged area image; if the degree of anisotropy is low and it is distributed in a mesh pattern, it is determined to be a drying shrinkage crack; if the degree of anisotropy is high and it is distributed in a radial pattern, it is determined to be erosion damage.

9. The method for inspecting dam maintenance gates according to claim 2, characterized in that: If the proportion of rust area on the surface of the inspection door (5) exceeds the rust area threshold dynamically adjusted based on its service life, or if the density of adjacent concrete cracks exceeds the corresponding service life-related threshold, then targeted maintenance suggestions will be generated respectively.

10. The method for inspecting dam maintenance gates according to claim 1, characterized in that: In S6, a segmented lifting strategy is adopted in the recycling stage. The pause time of each segment is dynamically extended according to the cumulative curvature of the segment and all previous curved segments, so that a longer cable relaxation time is given in the high curvature accumulation area, so that the flexible cable (20) naturally fits the hole wall of the vent (6) under the action of gravity.

Images