Cable bridge fatigue crack risk assessment method and system

Abstract

The invention relates to the field of bridge risk assessment, in particular to a cable bridge fatigue crack risk assessment method and system. The method comprises the following steps: carrying out omnibearing structure scanning on a cable bridge based on an unmanned aerial vehicle, and constructing a bridge image library; performing fatigue damage area distribution identification based on the deformation response data, and marking a potential fatigue damage area; performing crack texture visual analysis based on the potential fatigue damage area, and extracting crack damage characterization; and performing crack risk assessment according to the crack damage characterization, and outputting a bridge risk assessment report. According to the method, the comprehensiveness and accuracy of bridge crack risk assessment are improved, and the operation safety of the cable bridge is improved.

Description

Technical Field

[0001] This invention relates to the field of cable tray risk assessment, and in particular to a method and system for assessing fatigue crack risk in cable trays. Background Technology

[0002] During long-term operation, cable trays are subject to various factors such as load changes, equipment vibration, temperature and humidity fluctuations, and environmental corrosion. These factors gradually cause fatigue damage within the cable tray structure, leading to the initiation of microcracks at stress concentration points. As operating time increases, these cracks propagate and penetrate, reducing the cable tray's load-bearing capacity. In severe cases, this can cause cable tray breakage or collapse, resulting in damage to power or communication lines and affecting the safe and stable operation of the system. Therefore, early identification and risk assessment of fatigue cracks in cable trays are crucial for ensuring structural safety and extending their service life. Existing cable tray inspection and maintenance methods primarily rely on manual inspection, experience-based judgment, or localized non-destructive testing methods, such as visual inspection, impact testing, and limited sensor monitoring. These methods suffer from low efficiency, high subjectivity, and limited coverage in practical applications. Cracks are often difficult to detect in a timely manner, especially for cable tray structures located at heights, in confined spaces, or in concealed areas. Traditional detection methods typically rely on single test results, lacking continuous tracking and quantitative analysis of the crack development process. This makes it difficult to accurately assess the evolution trend and potential risks of fatigue cracks, and can easily lead to missed detections or misjudgments.

[0003] With the continuous expansion of cable tray scale and the increasing complexity of operating environments, traditional methods relying on manual labor and simple inspection tools are no longer sufficient to meet the demands for structural safety, real-time monitoring, and refined management of cable trays. Especially under long-term operating conditions, cable tray fatigue cracks are characterized by their high degree of concealment, slow development process, but severe consequences, necessitating a technological means to identify cracks and assess their risks over a wide range with high precision. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention proposes a method and system for assessing the risk of fatigue cracks in cable trays, thereby resolving at least one of the aforementioned technical issues.

[0005] To achieve the above objectives, the present invention provides a method for assessing the fatigue crack risk of cable trays, comprising the following steps: Step S1: Conduct a full-range structural scan of the cable tray using a drone to build a cable tray image library; Step S2: Identify the distribution of fatigue damage areas based on the deformation response data and mark potential fatigue damage areas; Step S3: Perform visual analysis of crack texture based on the potential fatigue damage area to extract crack damage characterization; Step S4: Conduct a crack risk assessment based on the crack damage characterization and output a cable tray risk assessment report.

[0006] This specification provides a cable tray fatigue crack risk assessment system for performing the cable tray fatigue crack risk assessment method described above, comprising: The image acquisition unit is used to perform omnidirectional structural scanning of cable trays based on drones and build a cable tray image library; The damage identification unit is used to identify the distribution of fatigue damage areas based on the deformation response data and mark potential fatigue damage areas. The visual analysis unit is used to perform visual analysis of crack texture based on potential fatigue damage areas and extract crack damage characterization. The risk assessment unit is used to assess crack risk based on crack damage characterization and output a cable tray risk assessment report.

[0007] The beneficial effects of this invention are as follows: By introducing drones to perform omnidirectional structural scanning of cable trays, the limitations of traditional manual inspections—limited coverage, single perspective, and difficulty in accessing high-altitude and concealed areas—are effectively solved. Drones can acquire high-resolution images from multiple angles, including the top, sides, and bottom of the cable tray, achieving comprehensive visual coverage of the overall structure and key nodes. By analyzing the cable tray deformation response data and identifying fatigue damage areas, key areas with concentrated stress and significant fatigue accumulation can be quickly located within the overall structure, narrowing the spatial range for crack detection, reducing the probability of false and missed detections, and improving the targeting and efficiency of crack identification. It effectively avoids interference from complex backgrounds, making crack detection more accurate and stable. By analyzing the crack texture features in the images, multi-dimensional damage characterization information reflecting crack morphology and development status can be extracted, such as crack length, width, direction, and texture continuity. Crack damage information is transformed into intuitive risk levels or risk indicators, giving cable tray fatigue crack risk a clear quantitative expression. The output cable tray risk assessment report can intuitively reflect the current operational risks and potential future failure trends of the cable tray, providing a basis for decision-making for the safety management of the cable tray throughout its entire life cycle, thereby significantly improving the practicality and engineering application value of cable tray fatigue crack risk assessment. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of the steps in the fatigue crack risk assessment method for cable trays according to the present invention; Figure 2 This is a detailed flowchart illustrating the implementation steps of step S1. Figure 3 This is a flowchart illustrating the detailed implementation steps of step S2. Detailed Implementation

[0009] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0010] This application provides a method and system for assessing the fatigue crack risk of cable trays. The implementing entities of the cable tray fatigue crack risk assessment method and system include, but are not limited to, mechanical equipment, data processing platforms, cloud server nodes, network upload devices, etc., which can be considered general computing nodes of this application. The data processing platform includes, but is not limited to, at least one of an audio-visual management system, an information management system, and a cloud data management system.

[0011] Please see Figures 1 to 3 This invention provides a method for assessing the fatigue crack risk of cable trays, comprising the following steps: Step S1: Conduct a full-range structural scan of the cable tray using a drone to build a cable tray image library; Step S2: Identify the distribution of fatigue damage areas based on the deformation response data and mark potential fatigue damage areas; Step S3: Perform visual analysis of crack texture based on the potential fatigue damage area to extract crack damage characterization; Step S4: Conduct a crack risk assessment based on the crack damage characterization and output a cable tray risk assessment report.

[0012] In the embodiments of the present invention, see Figure 1 The diagram below illustrates the steps of a cable tray fatigue crack risk assessment method according to the present invention. In this example, the steps of the cable tray fatigue crack risk assessment method include: Step S1: Conduct a full-range structural scan of the cable tray using a drone to build a cable tray image library; In this embodiment, a UAV flight path is designed based on the geometry and layout of the cable tray to achieve longitudinal, lateral, and diagonal surround scanning of the cable tray. The flight altitude is controlled within the range of 2.5 to 4 meters to ensure an image resolution of less than 0.3 mm / pixel, thereby capturing minute cracks and structural features. The camera is set to high resolution (50 megapixels or higher) with a focal length of approximately 24 mm to ensure image consistency from different viewing angles. During flight, repeated image acquisition is performed through preset waypoints, with a longitudinal overlap rate of no less than 75% and a lateral overlap rate of no less than 65%, ensuring coverage of the same area at different angles. The acquired images undergo distortion correction, brightness equalization, and contrast normalization processing, and are labeled with the acquisition time, flight parameters, and environmental conditions of each image to construct a complete cable tray image library.

[0013] Step S2: Identify the distribution of fatigue damage areas based on the deformation response data and mark potential fatigue damage areas; In this example, multi-point displacement tracking technology is used to extract two-dimensional displacement vectors of surface measurement points from a cable tray image database. Sub-pixel-level analysis is employed to improve the accuracy of the measurement point displacement and distinguish minute deformations. Multi-time-point tracking of the measurement point displacement paths yields continuous time-series displacement trajectories. Random environmental interference suppression methods, such as smoothing filtering and neighborhood consistency constraints, are used to eliminate abnormal noise interference. Next, spatial clustering and segmentation of the measurement point displacement paths are performed, grouping measurement points with similar displacement trends and spatial continuity into the same region to form deformation response data for the structural region. Based on the deformation attenuation rate and the degree of clustering of unstable deformation points, the degree of fatigue accumulation in the region is determined, and potential fatigue damage areas are marked. Each marked region includes its spatial location, the range of measurement points it covers, and the corresponding deformation response characteristics.

[0014] Step S3: Perform visual analysis of crack texture based on the potential fatigue damage area to extract crack damage characterization; In this embodiment, local image sequences of the damaged area are extracted from the cable tray image library. The cropping window size is controlled between 256×256 and 512×512 pixels to ensure complete representation of crack details. Super-resolution reconstruction of the local image sequences is performed, typically using a multi-frame fusion method to integrate minute image displacement information from the time series, improving spatial resolution and texture clarity, making crack edges and branch structures clearer. After reconstruction, texture enhancement and grayscale gradient analysis methods are used to identify candidate crack regions, and crack length, width, orientation, and branch morphology are quantitatively measured to form crack damage characterization data. The changes in crack characteristics are tracked through time series analysis to extract crack growth rate, acceleration, and cumulative deformation information.

[0015] Step S4: Conduct a crack risk assessment based on the crack damage characterization and output a cable tray risk assessment report.

[0016] In this embodiment, time-series analysis is performed on crack length, width, growth rate, and cumulative deformation to construct a crack change time axis. The single-cycle change and multi-cycle cumulative damage are calculated, and a fatigue deterioration trend index is generated. A rolling window prediction method is used to obtain crack propagation prediction data, including the joint probability distribution of crack initiation time, propagation rate, and size. Environmental factors are applied to the crack prediction data based on cable tray operating environment parameters, such as temperature cycling, vibration load, and corrosion degree, resulting in corrected prediction data. Based on the corrected prediction data, the expected remaining life before the crack reaches its critical size is calculated and compared with the design life and safety standards to identify high-risk areas and potential critical crack locations. The analysis results, crack distribution map, remaining life assessment, fatigue deterioration trend, and maintenance priority recommendations are integrated into a cable tray risk assessment report.

[0017] In this embodiment, see Figure 2The diagram below illustrates the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: A fixed image acquisition cycle is set, and the cable tray is scanned from all directions using a drone to extract multiple periodic image sets; Adaptive pixel filtering is applied to multiple periodic image sets to generate a filtered and optimized image set. Perform bridge structure feature point identification on the filtered and optimized image set, and mark the structural feature points; Temporal image registration is performed based on structural feature points to obtain image registration data; A cable tray image library is constructed by building a timestamp index based on image registration data.

[0018] In this embodiment, a stable and repeatable image acquisition cycle is determined based on the fatigue crack development characteristics of cable trays under long-term load, vibration, and environmental effects. The acquisition cycle is typically set to 7 days, 14 days, or 30 days, with shorter cycles used for areas with frequent vibration and significant temperature variations, and longer cycles used for structurally stable areas, to balance data continuity and storage efficiency. A UAV equipped with a high-resolution visible light imaging device performs multi-angle, full-coverage scanning of the entire cable tray structure. The camera resolution is set to above 50 megapixels, with a focal length maintained at approximately 24mm, ensuring that the spatial resolution of the image reaches within 0.3 mm / pixel within a flight altitude range of 2.5 meters to 4 meters, thus meeting the requirements for identifying minute cracks. The flight path is completed through a preset flight path, performing multiple scans around the cable tray longitudinally, laterally, and diagonally to ensure that the bottom, side walls, connection points, and weld areas of the cable tray are completely captured. During the shooting process, the image overlap rate is controlled, with a vertical overlap of no less than 75% and a horizontal overlap of no less than 65%. Brightness and contrast are standardized, and the overall brightness is normalized by statistically analyzing the grayscale distribution characteristics of images from different periods, ensuring visual consistency across images acquired at different times. A multi-scale filtering strategy is introduced at the pixel level, combining median filtering and edge-preserving filtering to optimize the image. The median filtering window size is typically set to 3×3 or 5×5 to eliminate isolated noise. Edge-preserving filtering controls the weights in the spatial and grayscale domains to preserve the edges of the cable tray, weld lines, and crack texture information while smoothing background noise. For bright areas on the metal cable tray surface caused by reflection, an adaptive threshold suppression method is used to adjust pixel weights, reducing interference from overexposed areas on crack identification.

[0019] Based on the structural form of the cable tray, key components were analyzed, with areas such as beam edges, column connections, bolt holes, weld intersections, and channel steel bends selected as key feature extraction targets. These areas exhibit minimal geometrical changes during long-term service, making them suitable as reference points for temporal alignment. During feature point recognition, image grayscale variations, gradient information, and local texture distribution were comprehensively utilized to detect corner points, edge intersections, and areas with significant texture. The feature response threshold was adaptively adjusted based on image resolution and noise level, ensuring both sufficient quantity and good stability of extracted feature points. To avoid misidentifying cracked areas as structural feature points, spatial distribution and geometric consistency constraints were introduced during recognition, and points with abnormal scale changes or significant positional drift were removed. The earliest acquired image was selected as the baseline reference, and subsequent images from other periods were sequentially matched with the baseline image for feature point identification. During matching, spatial distance between feature points, local descriptive similarity, and structural category consistency were comprehensively considered to select reliable matching point pairs. To ensure registration stability, the number of matching points was maintained at over 200 pairs, and abnormal matching points were removed. Based on the selected matching point pairs, geometric transformation relationships are constructed to correct for translation, rotation, and minute scale changes between images. To address potential local deformations of the cable tray during long-term service, fine adjustments are made to local areas based on overall alignment to ensure that structural features maintain consistent spatial positions across different nodes. After registration, the registration error is evaluated to ensure that the positional deviation of structural feature points in images across different periods is controlled within 1 pixel.

[0020] Each registered image is assigned a distinct time stamp, including the acquisition date, period number, and shooting sequence, and is associated with the corresponding structural feature point data. Using structural feature points as the core index unit, images of the same cable tray location at different nodes are concatenated to form a continuous time-series image set. The image database storage structure employs a hierarchical indexing approach, using cable tray number, structural location, and timestamp as multi-level indexing conditions to achieve rapid retrieval and comparative analysis. Through this method, images at the same location at different nodes can be directly aligned at the pixel level and compared for crack variations.

[0021] In this embodiment, see Figure 3 The diagram below illustrates the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Multi-point displacement tracking was performed using a cable tray image library to obtain deformation response data for different structural regions; Based on the deformation response data, a deformation response time sequence consistency analysis is performed to obtain the repeatability characteristics of the deformation mode; Based on the repeating characteristics, the attenuation trend is calculated, and the deformation attenuation rate of different regions is output. Unstable deformation is detected based on the deformation attenuation rate, and deformation instability points are marked. Based on the distribution of the deformation instability points, potential fatigue damage areas are marked.

[0022] In this embodiment, images that have completed time-series registration in the cable tray image library are used as input, and structural feature points that have been identified and are stable in the early stages are selected as displacement tracking objects. These feature points cover the cable tray beams, columns, connection nodes, weld areas, and load-bearing parts. A certain number of tracking points are evenly distributed in each structural area, with 30 to 80 tracking points typically set in a single area to ensure the completeness of deformation description. Using timestamps as indexes, the spatial positions of the same feature point at different nodes are compared one by one, and its displacement change in the plane coordinate system is calculated. The displacement calculation accuracy is controlled at the sub-pixel level. By interpolating and suppressing errors in the neighborhood of the feature points, it is ensured that small deformations can be accurately captured. To avoid single-point anomalies interfering with the overall analysis, a statistical summarization method is introduced at the regional level. The displacement results of all tracking points in the same region are weighted and fused to form a deformation response data sequence for that region at each time node.

[0023] Deformation response data from each region are normalized chronologically to eliminate the influence of differences in shooting scale and initial conditions, ensuring comparability of data across different segments. The deformation response sequence is segmented into fixed time windows, typically 3 to 5 acquisition cycles long, to extract local deformation characteristics. Within each time window, key indicators such as deformation amplitude, direction of change, and rate of change are analyzed and combined to form a regional deformation pattern description. By comparing the similarity of deformation patterns in adjacent and non-adjacent time windows, patterns that repeatedly appear across multiple time periods are identified. When a deformation pattern appears a certain number of times in the overall time series reaching a set threshold (e.g., exceeding 60%), it is determined to be a recurring deformation feature of that region.

[0024] The frequency of deformation patterns occurring in each time window is used as a repeatability index, with the frequency of repetitive features in the initial stage serving as a reference benchmark. The repeatability index is continuously statistically analyzed along the time axis, recording its changes in subsequent cycles. As the cable tray structure gradually experiences fatigue accumulation, the previously stable and repeating deformation patterns exhibit characteristics such as decreased frequency, increased deformation amplitude, or enhanced fluctuations. By fitting and analyzing the repeatability change curve, the rate of decrease per unit time is calculated to obtain the deformation attenuation rate. Smoothing is introduced during the attenuation rate calculation to avoid interference from short-term fluctuations in trend judgment. The attenuation rate is typically divided into three intervals: low attenuation, medium attenuation, and high attenuation. The high attenuation region indicates a significant decrease in the structural response stability of that area. An attenuation rate threshold is set; when the deformation attenuation rate in a certain region consistently exceeds this threshold, the deformation pattern in that region is considered to be transitioning from a stable to an unstable state. Further detailed analysis is conducted within each region, comparing the displacement response of each tracking point within the region with the overall trend of the region, focusing on points where the displacement change amplitude significantly deviates from the average level. When a tracking point exhibits abrupt displacement changes, directional reversals, or high-frequency fluctuations across multiple consecutive time windows, and this pattern is inconsistent with the dominant deformation mode in the region, it is marked as a deformation instability point. To avoid misjudgments due to accidental interference, a persistent judgment condition is set for deformation instability points, requiring that abnormal deformation recur within at least two adjacent periods. All deformation instability points are mapped back to their spatial locations on the cable tray structure, categorized and statistically analyzed according to structural parts, and their concentration in different regions is examined. When multiple deformation instability points exhibit significant spatial clustering, concentrated near welds, connection nodes, or locations of abrupt cross-sectional changes, it indicates a high risk of fatigue damage in that region. Based on the deformation attenuation rate corresponding to each region, a comprehensive evaluation of the instability point clustering areas is conducted. Regions with high attenuation rates and dense instability points are preferentially marked as potential fatigue damage areas. To improve identification reliability, the region boundaries are extended, including structural units within a certain range around the instability points in the risk area marking to cover potential paths for crack initiation and propagation.

[0025] In this embodiment, the specific steps for obtaining deformation response data of different structural regions by performing multi-point displacement tracking based on the cable tray image library are as follows: Based on the cable tray image library, surface measurement points are marked for each image, and multiple measurement points on the cable tray surface are extracted. Subpixel-level analysis was performed on multiple measuring points on the surface of the cable tray to extract two-dimensional displacement vectors for different measuring points; Random environmental interference is identified on the two-dimensional displacement vector, and interference suppression processing is performed to obtain the interference-suppressed displacement vector; Multi-time point optical flow tracing is performed on the interference suppression displacement vector to extract the displacement path at different measurement points; Spatial clustering and segmentation are performed based on the displacement path to obtain deformation response data for different structural regions.

[0026] In this embodiment, based on the structural form and stress characteristics of the cable tray, the surface of the cable tray is divided into regions, with the web of the crossbeam, sidewalls, bottom support arms, connection nodes, and areas near welds designated as key areas for measuring points. In each image, based on the spatial registration results, the distribution of measuring points at the same structural locations is kept consistent. The selection of measuring points follows the principle of combining uniform coverage with focused density. Basic measuring points are laid out within the overall area at a fixed pixel spacing, typically controlled within the range of 20 to 40 pixels. The distribution of measuring points is appropriately densified in welds, hole edges, and stress concentration areas, with the spacing reduced to 10 to 15 pixels to improve the ability to perceive local deformation. During the measuring point marking process, image edges, occluded areas, and areas with unstable textures are filtered out to prevent measuring points from falling in locations where reliable tracking is not possible. Each image forms a stable set of multiple surface measuring points, and each measuring point is assigned a unique number and its initial position in the image coordinate system. Using two adjacent registered images in the time series as comparison objects, the positional changes of the same numbered measuring point in the preceding and following images are calculated. To overcome the limitations of pixel-level accuracy, an interpolation analysis method is introduced within the neighborhood of the measurement point to continuously process the grayscale distribution, thereby achieving sub-pixel-level positioning. The neighborhood window size is typically set to 9×9 or 11×11 pixels to ensure local texture integrity while avoiding excessive background interference. By analyzing the grayscale gradient changes within the neighborhood of the measurement point, the displacement components in the horizontal and vertical directions are calculated, forming a two-dimensional displacement vector. The displacement resolution accuracy can be controlled within 0.05 pixels, effectively capturing the minute elastic deformations and cumulative deformations generated by the cable tray during long-term service. To improve the stability of the results, consistency verification is performed on the displacement vectors at continuous time points, eliminating results with abrupt changes in direction or abnormal amplitude. Due to the unavoidable influence of wind disturbances, lighting changes, and local reflections during UAV data acquisition, some displacement vectors may exhibit strong randomness, irregular direction, or abnormal amplitude fluctuations. Statistical analysis is performed on the displacement vector sequence of each measurement point to calculate its mean, variance, and variation amplitude within a certain time window. When displacement changes frequently exhibit high-frequency oscillations or sudden increases, and are significantly inconsistent with the overall trend of surrounding measuring points, they are identified as interfering candidate data. A spatial consistency constraint is introduced, comparing the displacement of a measuring point with that of its neighboring points. If the displacement direction and amplitude of a particular measuring point deviate significantly from those of most points in the neighborhood, it is further confirmed that it is affected by random interference. For the identified interfering displacements, smoothing filtering and weight redistribution are used to suppress them, reducing the impact of anomalous components on the overall analysis and preserving the main trend of the true structural deformation.

[0027] Starting from the initial position of the measuring point at the initial time node, and combining the displacement vectors of subsequent time nodes, the motion trajectory of the measuring point on the time axis is gradually accumulated and calculated. During the tracking process, the continuity of the displacement path is constrained to avoid path breakage or deviation due to data fluctuations at individual time nodes. To ensure path smoothness, a smoothing process in the time dimension is introduced, enabling the displacement path to accurately reflect the gradual deformation process of the cable tray structure. Each displacement path consists of displacement points at multiple time nodes, clearly describing the movement direction, cumulative displacement, and trend of the measuring point within its service life. Cases of significant bending, acceleration, or reversal in the displacement path are individually marked for subsequent structural response analysis. The displacement paths of all measuring points are mapped back to the spatial location of the cable tray, and the direction, amplitude, and cumulative displacement characteristics of the paths are comprehensively analyzed. By comparing the similarity between the displacement paths of different measuring points, measuring points with similar motion trends and adjacent spatial locations are grouped into the same cluster unit. The clustering process considers spatial distance and displacement characteristic similarity, ensuring that the partitioning results both conform to the structural geometric distribution and reflect the actual deformation behavior. The clustering scale is adjusted according to the dimensions of the cable tray structure, and each cluster is usually controlled to cover a structural component or its key local area. After the clustering is completed, the displacement paths of the measuring points in each cluster are summarized and analyzed to extract the average displacement response, maximum displacement amplitude and variation trend of the region at different nodes, forming regional deformation response data.

[0028] In this embodiment, the specific steps for performing spatial clustering and segmentation based on the displacement path to obtain deformation response data for different structural regions are as follows: Calculate the spatial coordinates of multiple measuring points on the surface of the cable tray; The cable tray position is identified based on the spatial coordinates to obtain the structural regions of different measuring points; Frequency domain transformation is performed on the displacement paths of different measuring points to identify the dominant deformation frequency components and harmonic characteristics; Based on the structural regions, spatial clustering and segmentation of the dominant deformation frequency components and harmonic characteristics are performed to obtain deformation response data for different structural regions.

[0029] In this embodiment, based on the registered images in the cable tray image library, and combined with the camera intrinsic parameters information from the drone shooting, the image coordinates are geometrically corrected. Camera intrinsic parameters include parameters such as focal length, principal point position, and pixel size; these parameters are used to eliminate lens distortion and restore true proportional relationships. Based on the relative altitude and viewing angle information of the drone during the shooting process, the measurement point positions in the image are converted into three-dimensional spatial coordinates corresponding to the cable tray surface. During the spatial coordinate calculation, it is assumed that the local surface of the cable tray is approximately planar within the neighborhood of the measurement point, and planar constraints are used to reduce errors caused by depth uncertainty. To improve coordinate stability, the spatial coordinates calculated at the same measurement point at multiple time points are checked for consistency. When the deviation exceeds a set threshold (e.g., 1 to 2 mm), outliers are corrected or removed. Combining the structural design characteristics of the cable tray, the overall cable tray is spatially divided into multiple structural regions, including beam areas, column areas, connection node areas, weld concentration areas, and edge free segments. Each region has a clearly defined geometric range and functional attributes in space. The spatial coordinates of the measurement points are matched with the boundaries of predefined structural regions, and the structural region to which the measurement point belongs is automatically determined based on its spatial location. For measurement points located at the boundary of regions, distance weights and neighborhood consistency rules are introduced to avoid misassignment. To ensure the stability of the region identification results, the region assignment of the measurement points is checked for consistency over time, requiring the same measurement point to maintain the same region label at adjacent time points unless its spatial location changes significantly.

[0030] The displacement paths of measuring points at continuous time points are compiled into a time series signal, reflecting the periodic response and random fluctuation characteristics of the measuring points during their service life. This time series is then subjected to frequency domain transformation, converting the time-domain displacement signal into a frequency-domain expression to extract the energy distribution of different frequency components. In the frequency domain results, frequency components with high amplitude and stable existence are identified as dominant deformation frequencies, typically corresponding to the inherent response characteristics of the cable tray under load or environmental excitation. Frequency peaks appearing near integer multiples of the dominant frequency are identified as harmonic features, used to reflect the structural response characteristics under nonlinear states. The frequency resolution is set according to the time series length, typically controlled within the range of 0.01 to 0.05 Hz to balance frequency accuracy and computational stability. Frequency domain characteristic parameters of all measuring points within the same structural region are summarized, including dominant frequency values, frequency amplitudes, and harmonic distribution characteristics. Similarity analysis is performed on these parameters, grouping measuring points with similar frequency characteristics and harmonic structures into the same cluster unit. During clustering, the spatial proximity of measurement points is considered to ensure the geometrical continuity of the clustering results and avoid unreasonable divisions that are spatially discrete but have similar frequencies. Statistical analysis of the clustering results is performed to extract representative deformation frequency characteristics within each structural region, including the dominant frequency range, harmonic energy distribution, and frequency stability indices.

[0031] In this embodiment, the specific steps of step S3 are as follows: Based on the potential fatigue damage area, local image extraction is performed on the cable tray image library to extract the local image sequence of the damage area; Super-resolution reconstruction is performed on the local image sequence to obtain a super-resolution image sequence; Visual analysis of crack texture is performed on super-resolution image sequences to extract crack damage characteristics.

[0032] In this embodiment, the location of the corresponding region is determined in the time-registered cable tray image based on the spatial boundary information of the potential fatigue damage area. This region is typically located near the weld, connection node, or location of structural cross-section changes, and its spatial range is determined by the clustering distribution of early deformation instability points. Centered on this region, a fixed-size local image block is cropped from the image corresponding to each time node. The cropping window size is set according to the original image resolution, generally controlled within the range of 256×256 to 512×512 pixels, to reduce redundant background information while ensuring the complete presentation of crack details. To avoid information loss caused by boundary cropping, the cropping window is appropriately expanded so that the damage area is located at the center of the local image. By maintaining the consistency of the cropping window in the time dimension, a sequence of local images arranged in chronological order is formed. The local image sequence is preprocessed, including grayscale normalization, noise suppression, and contrast enhancement, to ensure consistency in brightness and texture at each time node. Utilizing the small displacements and information redundancy between multiple frames in the time series, multi-frame fusion super-resolution reconstruction is performed on the local images. The reconstruction magnification is typically set to 2x or 4x to effectively amplify the edges of fine cracks that would otherwise be difficult to distinguish. During reconstruction, the texture continuity of the edge regions is carefully constrained by assigning higher weights to areas with significant gradient changes to prevent the crack outlines from being smoothed or blurred during magnification. To ensure the stability of the super-resolution results, consistency checks are performed on the reconstructed images at consecutive time points to ensure that the texture enhancement originates from genuine structural features rather than reconstruction artifacts.

[0033] Texture enhancement processing is performed on the super-resolution image to highlight regions of abrupt grayscale changes and linear structures, creating a clear contrast between the crack and the background surface texture. Frame-by-frame analysis of candidate crack regions extracts geometric features such as crack length, width, direction, and branching morphology, with crack width measurement accuracy controlled to the order of 0.1 mm. Further analysis of the texture distribution characteristics of the crack region, addressing crack edge roughness and grayscale inhomogeneity, is used to characterize crack propagation and fatigue accumulation properties. By comparing crack feature changes over time, crack growth rate and morphological evolution trends are extracted, distinguishing between stable cracks and rapidly propagating cracks.

[0034] In this embodiment, the specific steps of step S4 are as follows: Damage accumulation rate analysis was performed on crack damage characterization to obtain fatigue deterioration trend indicators; A rolling window prediction is performed based on fatigue deterioration trend indicators to obtain crack propagation prediction data; the crack propagation prediction data includes the joint probability distribution of crack initiation time, propagation rate, and size; Extract the environmental parameters of the cable tray operating conditions; these include temperature cycling, vibration load, and corrosion degree. Based on the aforementioned cable tray operating environment parameters, the crack propagation prediction data is corrected using an environmental coefficient to obtain corrected prediction data. The expected remaining lifetime value when the crack reaches the critical size is calculated based on the corrected prediction data; Based on the expected remaining life, a crack risk assessment is performed, and a cable tray risk assessment report is output.

[0035] In this embodiment, crack length, width, and crack growth rate data are organized into a time series, and the cumulative crack growth per unit time is calculated for each measuring point or region. Fitting methods, such as exponential fitting, moving average, or piecewise linear regression, are used to extract the trend of the cumulative crack rate and mark the time periods of sudden rate increases as key stages of fatigue deterioration. During the fitting process, the time window length and sampling period are controlled, typically consistent with the image acquisition period, such as 7 or 14 days, to ensure that the trend analysis takes into account both short-term fluctuations and long-term development characteristics. The fatigue deterioration trend curve is piecewise fitted using a rolling time window as the unit. By analyzing the local changes in crack growth rate within each window, the possible crack propagation in the next time period is predicted. The window length is typically set to 3 to 5 acquisition periods to balance the stability of the prediction with the sensitivity to short-term accelerated crack changes. During the prediction process, the fitting results of each time window are accumulated and superimposed to form a continuous time series prediction, outputting crack propagation prediction data. The prediction data includes the joint probability distribution of crack initiation time, propagation rate, and crack size, and statistical methods are used to estimate the probability of different crack evolution paths.

[0036] To improve the reliability of crack prediction, key parameters of the cable tray's operating environment need to be extracted, including temperature cycling, vibration load, and corrosion level. Temperature cycling is obtained by monitoring the diurnal temperature range and seasonal variations of the environment where the cable tray is located, reflecting the impact of thermal expansion and contraction on crack development. Vibration load is obtained from accelerometer data or historical vibration spectra of the structure where the cable tray is located, used to quantify mechanical fatigue effects. Corrosion level is obtained through assessment of environmental humidity, pollution index, and material surface condition, used to correct for the impact of decreased material strength on crack propagation. Based on the temperature cycling intensity, vibration load amplitude, and corrosion level, the corresponding environmental acceleration factor is calculated, reflecting the enhancement or mitigation effect on crack propagation rate. Then, the environmental acceleration factor is applied to the joint probability distribution of crack propagation prediction data to adjust crack initiation time, propagation rate, and size. High-temperature and high-humidity corrosive environments increase the probability of crack propagation and accelerate initiation, while low-vibration environments may delay crack growth. The corrected prediction data can more realistically reflect the crack evolution trend of the cable tray under actual service conditions.

[0037] Based on the crack critical size standard or design safety limit, the time point at which each predicted path reaches the critical state is identified. Weighted statistics are performed on different predicted paths and their probability distributions to calculate the expected time for the crack to reach the critical size, yielding the expected remaining life value. The calculation considers the initial crack size, current propagation rate, and acceleration changes to ensure the expected remaining life value reflects the dynamic characteristics of crack growth. Risk level thresholds are set, such as high risk, medium risk, and low risk, based on the proportion of the expected remaining life value to the design service life. The remaining life of each key area of ​​the cable tray is statistically analyzed and summarized to identify potential risk concentration areas in the overall structure. Combining crack propagation rate, acceleration, and cumulative damage, qualitative and quantitative analyses of the local or overall fatigue state of the cable tray are conducted. The aforementioned crack evolution analysis, remaining life calculation, and risk assessment results are integrated into a cable tray risk assessment report. The report includes a crack distribution map, crack propagation prediction curves, cumulative damage statistics, an expected remaining life value table, and regional risk level indicators. It provides fatigue deterioration trends, potential critical time points, and maintenance priority recommendations for different areas of the cable tray.

[0038] In this embodiment, the specific steps for performing damage accumulation rate analysis on crack damage characterization to obtain fatigue deterioration trend indicators are as follows: The crack damage characterization is fitted with a time-series periodic variation to generate a crack change time axis; Calculate the crack deformation growth rate and acceleration parameters based on the crack change time axis; Based on a fixed image acquisition cycle, the crack deformation growth rate and acceleration parameters are calculated in a single cycle to obtain the single-cycle change. The cumulative damage is obtained by performing linear cumulative analysis over multiple periods based on the single-period change. Damage accumulation rate analysis was performed on the accumulated damage to obtain fatigue deterioration trend indicators.

[0039] In this embodiment, the characterization indicators such as crack length, width, and branch propagation are organized into a continuous time series according to the time sequence of the cable tray image library. A curve fitting method is used for this time series; commonly used fitting models include polynomial fitting, exponential function fitting, and piecewise linear fitting to capture the slow propagation and local acceleration trends of cracks over time. During the fitting process, noise and outliers between data points are considered, and the fitting results are optimized using weighted least squares to ensure that the fitted curve can smooth the overall trend while reflecting short-cycle crack abrupt changes. The horizontal axis of the fitted curve corresponds to the acquisition timestamp, and the vertical axis corresponds to the crack characterization value, forming a complete crack change time axis. The crack length or width changes at consecutive time nodes on the time axis are differentially calculated to obtain the crack growth rate within each time interval. The time interval is usually consistent with the image acquisition period, such as 7 days or 14 days, to ensure that the rate calculation reflects the actual crack development changes. The rate sequence is differentially processed again to obtain the crack growth acceleration parameter, which is used to describe the trend of crack growth rate over time. Abnormal fluctuations are smoothed during the calculation process, for example, by using moving average or weighted filtering methods, so that the rate and acceleration indicators can more accurately reflect the true development characteristics of the crack.

[0040] The crack growth rate curve is segmented according to the acquisition period, with each slice corresponding to the incremental change of the crack within one acquisition period. The single-period change includes the increment of crack length or width, the rate change amplitude, and the acceleration change amplitude, reflecting the instantaneous dynamic characteristics of crack development within that period. To ensure calculation accuracy, the data within the period is interpolated to make the crack change curve continuous and smooth in the time dimension. By statistically analyzing and normalizing the single-period change of the crack at each measuring point, the local dynamic characteristics of crack growth in that period can be obtained. Multi-period cumulative analysis of crack damage is performed to quantify the overall cumulative amount of potential fatigue damage to the cable tray. During the accumulation process, the increments of crack length or width in each period are linearly superimposed in chronological order to obtain the total crack propagation amount across multiple acquisition periods. The single-period crack growth rate and acceleration are superimposed to generate multi-period cumulative rate curves and acceleration change curves. The cumulative analysis can be performed at different measuring points or summarized within a structural region to extract regional-level cumulative damage characteristics. During the calculation process, outlier correction is performed on the accumulated results to remove erroneous growth caused by image noise or local measurement point drift. The accumulated damage is segmented along the time axis, and the damage accumulation rate for each time period is calculated, i.e., the rate of change of crack propagation per unit time. The accumulation rate curve is fitted and trend analyzed using linear regression or exponential fitting methods to extract the overall development trend and local acceleration stages. The analysis can also incorporate crack acceleration indices to further identify accelerated fatigue damage stages and potential critical states.

[0041] In this embodiment, a cable tray fatigue crack risk assessment system is provided for performing the cable tray fatigue crack risk assessment method described above, including: The image acquisition unit is used to perform omnidirectional structural scanning of cable trays based on drones and build a cable tray image library; The damage identification unit is used to identify the distribution of fatigue damage areas based on the deformation response data and mark potential fatigue damage areas. The visual analysis unit is used to perform visual analysis of crack texture based on potential fatigue damage areas and extract crack damage characterization. The risk assessment unit is used to assess crack risk based on crack damage characterization and output a cable tray risk assessment report.

[0042] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, 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 the equivalents of the application are intended to be included within the invention.

[0043] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein are implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for assessing the risk of fatigue cracks in cable trays, characterized in that, Includes the following steps: Step S1: Conduct a full-range structural scan of the cable tray using a drone to build a cable tray image library; Step S2: Identify the distribution of fatigue damage areas based on the deformation response data and mark potential fatigue damage areas; Step S3: Perform visual analysis of crack texture based on the potential fatigue damage area to extract crack damage characterization; Step S4: Conduct a crack risk assessment based on the crack damage characterization and output a cable tray risk assessment report.

2. The method for assessing the fatigue crack risk of cable trays according to claim 1, characterized in that, The specific steps of step S1 are as follows: A fixed image acquisition cycle is set, and the cable tray is scanned from all directions using a drone to extract multiple periodic image sets; Adaptive pixel filtering is applied to multiple periodic image sets to generate a filtered and optimized image set. Perform bridge structure feature point identification on the filtered and optimized image set, and mark the structural feature points; Temporal image registration is performed based on structural feature points to obtain image registration data; A cable tray image library is constructed by building a timestamp index based on image registration data.

3. The method for assessing the fatigue crack risk of cable trays according to claim 1, characterized in that, The specific steps of step S2 are as follows: Multi-point displacement tracking was performed using a cable tray image library to obtain deformation response data for different structural regions; Based on the deformation response data, a deformation response time sequence consistency analysis is performed to obtain the repeatability characteristics of the deformation mode; Based on the repeating characteristics, the attenuation trend is calculated, and the deformation attenuation rate of different regions is output. Unstable deformation is detected based on the deformation attenuation rate, and deformation instability points are marked. Based on the distribution of the deformation instability points, potential fatigue damage areas are marked.

4. The fatigue crack risk assessment method for cable trays according to claim 3, characterized in that, The specific steps for obtaining deformation response data of different structural regions by performing multi-point displacement tracking based on the cable tray image library are as follows: Based on the cable tray image library, surface measurement points are marked for each image, and multiple measurement points on the cable tray surface are extracted. Subpixel-level analysis was performed on multiple measuring points on the surface of the cable tray to extract two-dimensional displacement vectors for different measuring points; Random environmental interference is identified on the two-dimensional displacement vector, and interference suppression processing is performed to obtain the interference-suppressed displacement vector; Multi-time point optical flow tracing is performed on the interference suppression displacement vector to extract the displacement path at different measurement points; Spatial clustering and segmentation are performed based on the displacement path to obtain deformation response data for different structural regions.

5. The method for assessing the fatigue crack risk of cable trays according to claim 4, characterized in that, The specific steps for spatial clustering and segmentation based on the displacement path to obtain deformation response data for different structural regions are as follows: Calculate the spatial coordinates of multiple measuring points on the surface of the cable tray; The cable tray position is identified based on the spatial coordinates to obtain the structural regions of different measuring points; Frequency domain transformation is performed on the displacement paths of different measuring points to identify the dominant deformation frequency components and harmonic characteristics; Based on the structural regions, spatial clustering and segmentation of the dominant deformation frequency components and harmonic characteristics are performed to obtain deformation response data for different structural regions.

6. The method for assessing the fatigue crack risk of cable trays according to claim 1, characterized in that, The specific steps of step S3 are as follows: Based on the potential fatigue damage area, local image extraction is performed on the cable tray image library to extract the local image sequence of the damage area; Super-resolution reconstruction is performed on the local image sequence to obtain a super-resolution image sequence; Visual analysis of crack texture is performed on super-resolution image sequences to extract crack damage characteristics.

7. The method for assessing the fatigue crack risk of cable trays according to claim 6, characterized in that, The crack damage characterization includes linear initiation fatigue cracks, bifurcation propagation fatigue cracks, network microcrack clusters, and fatigue cracks along the weld line.

8. The method for assessing the fatigue crack risk of cable trays according to claim 1, characterized in that, The specific steps of step S4 are as follows: Damage accumulation rate analysis was performed on crack damage characterization to obtain fatigue deterioration trend indicators; A rolling window prediction is performed based on fatigue deterioration trend indicators to obtain crack propagation prediction data; the crack propagation prediction data includes the joint probability distribution of crack initiation time, propagation rate, and size; Extract the environmental parameters of the cable tray operating conditions; these include temperature cycling, vibration load, and corrosion degree. Based on the aforementioned cable tray operating environment parameters, the crack propagation prediction data is corrected using an environmental coefficient to obtain corrected prediction data. The expected remaining lifetime value when the crack reaches the critical size is calculated based on the corrected prediction data; Based on the expected remaining life, a crack risk assessment is performed, and a cable tray risk assessment report is output.

9. The method for assessing the fatigue crack risk of cable trays according to claim 1, characterized in that, The specific steps for analyzing the damage accumulation rate of crack damage characterization to obtain fatigue deterioration trend indicators are as follows: The crack damage characterization is fitted with a time-series periodic variation to generate a crack change time axis; Calculate the crack deformation growth rate and acceleration parameters based on the crack change time axis; Based on a fixed image acquisition cycle, the crack deformation growth rate and acceleration parameters are calculated in a single cycle to obtain the single-cycle change. The cumulative damage is obtained by performing linear cumulative analysis over multiple periods based on the single-period change. Damage accumulation rate analysis was performed on the accumulated damage to obtain fatigue deterioration trend indicators.

10. A fatigue crack risk assessment system for cable trays, characterized in that, The method for performing the cable tray fatigue crack risk assessment method as described in claim 1 includes: The image acquisition unit is used to perform omnidirectional structural scanning of cable trays based on drones and build a cable tray image library; The damage identification unit is used to identify the distribution of fatigue damage areas based on the deformation response data and mark potential fatigue damage areas. The visual analysis unit is used to perform visual analysis of crack texture based on potential fatigue damage areas and extract crack damage characterization. The risk assessment unit is used to assess crack risk based on crack damage characterization and output a cable tray risk assessment report.

Images