Concrete structure crack full life cycle monitoring method and system

Abstract

The invention relates to the technical field of crack monitoring, in particular to a concrete structure crack full-life-cycle monitoring method and system, and the method comprises the following steps: arranging sound wave points to obtain a stress change to generate an initial signal, extracting an image feature recognition extension path, analyzing a size acceleration to generate an extension identifier, and screening an active cycle at intersection time to generate an evolution fragment. And classifying crack behaviors according to displacement and width trends to generate an annotation set. According to the method, stress evolution is reflected based on interval transit time, a crack expansion trend is extracted in combination with an image feature point track, an active state is identified by superimposing crack size speed increase synchronism, a high-frequency change fragment is positioned through multi-index time intersection, and crack behavior types are divided according to a displacement direction and a width increase trend. Continuous tracking and classified marking of the whole process from crack starting to evolution are achieved, the response capacity to abnormal crack behaviors is enhanced, and the precision and timeliness of structural safety state recognition are improved.

Description

Technical Field

[0001] This invention relates to the field of crack monitoring technology, and in particular to a method and system for monitoring cracks in concrete structures throughout their entire life cycle. Background Technology

[0002] The field of crack monitoring technology mainly involves technologies for identifying, tracking, and assessing cracks that occur in building structures, especially concrete structures, during their service life. Core aspects include crack detection, recording and analysis of crack morphology and propagation, and assessment and early warning of crack development trends. This technology encompasses various monitoring methods, including physical sensor-based monitoring, image recognition, and acoustic detection, and is widely used in the safety monitoring and maintenance management of critical infrastructure such as bridges, tunnels, dams, and high-rise buildings. Its systematic nature is reflected in the closed-loop management process of continuous monitoring, data acquisition, analysis, and processing of the structural safety status throughout its entire life cycle. Traditional full-life-cycle monitoring methods for concrete structure cracks refer to methods for monitoring the crack behavior that may occur in concrete structures from construction to operation. The key technical issue addressed by this method is how to achieve the perception and evolution recording of cracks from their initial appearance to propagation and eventual structural failure. Traditional methods for monitoring the entire life cycle of cracks in concrete structures typically employ specific techniques such as distributed fiber optic sensing, digital image correlation (DIC), and acoustic emission (AE) technology. These methods involve deploying optical fibers on or inside the structure, acquiring image sequences, or receiving acoustic emission signals to monitor and analyze the location, length, width, propagation path, and development rate of cracks.

[0003] Existing technologies primarily rely on single data sources, making it difficult to establish stress evolution processes from a temporal perspective using acoustic emission signals. Image contour change analysis focuses on static morphological comparisons and lacks methods for extracting dynamic offset trends in crack paths. This makes it difficult to achieve multi-dimensional linkage perception when crack behavior enters the early active stage. The lack of a time synchronization mechanism among monitoring indicators leads to unclear identification of active cycles. Crack behavior classification is based on single-point static results, failing to reflect the trend evolution characteristics under time series, resulting in delayed response to local structural anomalies. This can easily create monitoring blind spots, especially in complex areas such as bridge-tunnel junctions or high-stress zones, causing one-sided and lagging structural risk assessments. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a method for monitoring the entire life cycle of cracks in concrete structures, comprising the following steps:

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for monitoring the entire life cycle of cracks in concrete structures, comprising the following steps:

[0006] S1: Set up acoustic points at the downstream pier nodes of the dam, measure the propagation time to establish a time series, calculate the rate of change of sound velocity, obtain the value of concrete stress change, compare the stress curve with the fracture reference curve, and generate the crack initiation area signal.

[0007] S2: Obtain the inner wall image sequence of the crack initiation region signal, extract boundary feature points, calculate the Euclidean distance and angle change of the differential time, and if the direction is continuous and the distance increases, generate the crack outline expansion path.

[0008] S3: Based on the crack profile expansion path, extract the crack length and width sequence at the bottom edge of the main span, calculate the growth rate, determine whether the two increase synchronously and the difference decreases, and generate a crack synchronous expansion stage identifier.

[0009] S4: Call the crack initiation region signal, the crack outline expansion path and the crack synchronous expansion stage identifier, perform intersection operation based on the timestamps of the three sets of results, filter the cycles that meet the active threshold, and generate crack evolution active segment records;

[0010] S5: Based on the records of active crack evolution segments, analyze the changes in crack displacement, direction and width in time series, classify those with consistent direction and continuously increasing width into the extended class, and the rest into the stable class, and generate a crack evolution classification label set.

[0011] As a further aspect of the present invention, the crack initiation region signal includes the upper limit threshold of sound velocity change, the deviation amplitude of stress change, and the critical value triggering determination result; the crack profile expansion path includes the feature point offset direction, the distance change result between feature points, and the path stability determination result; the crack synchronous expansion stage identifier includes the crack length growth rate, the crack width growth rate, and the growth rate difference change trend; the crack evolution active segment record includes the multi-index cross time period, the crack state activity level, and the high-frequency evolution time segment; and the crack evolution classification annotation set includes the crack expansion direction, the crack width growth trend, and the behavioral continuity characteristics.

[0012] As a further aspect of the present invention, the synchronous increase and the narrowing difference are defined as the main span bottom edge crack length and width increasing together over time, and the difference between their growth gradually decreasing.

[0013] As a further aspect of the present invention, the extended class is defined as one whose crack direction is consistent with that of the stable class and whose width continues to increase; otherwise, it is defined as a stable class.

[0014] As a further aspect of the present invention, the specific steps of S1 are as follows:

[0015] S101: Set up sound wave transmitting and receiving points at the downstream pier nodes of the concrete dam, obtain the start and arrival times of sound wave propagation between nodes, establish node pairs based on the relationship between propagation duration and distance between nodes, and generate a sound wave propagation time series group.

[0016] S102: Based on the sound wave propagation time series group, call the node pairs to change the propagation time, calculate the ratio of the propagation time per unit time to the node distance, aggregate the results of the region, and obtain the sound speed change rate distribution map;

[0017] S103: Based on the sound velocity change rate distribution map and the structural response coefficient of the nodal region, calculate the force change per unit area according to the time series, analyze the force function of the nodal region, and generate the crack initiation region signal.

[0018] As a further aspect of the present invention, the specific steps of S2 are as follows:

[0019] S201: Obtain a sequence of differentiated time period images on the inner wall surface of the tunnel covered by the signal in the crack initiation area, extract all curvature reversal points and abrupt change direction points of the boundary contour curve in each image, combine them to form a set of feature point coordinates for the corresponding time period, and generate a set of feature points for the time period.

[0020] S202: Based on the set of feature points in the time period, calculate the Euclidean distance and the change in the included angle between corresponding feature points in adjacent time periods, arrange them sequentially to construct the temporal offset path of the feature points, and compare the changes in direction and distance of the path according to time to obtain a continuous offset direction sequence;

[0021] S203: Based on the continuous offset direction sequence, determine whether the path direction is continuous and consistent and whether the corresponding distance has an increasing trend, filter the path segments that meet the stable offset threshold, locate the contour boundary change trend of the offset concentration area, and generate the crack contour expansion path.

[0022] As a further aspect of the present invention, the specific steps of S3 are as follows:

[0023] S301: Call the bridge main span bottom edge area covered by the crack outline expansion path, obtain the crack length data and width data in the current period area, perform difference calculation on the crack length data of two adjacent periods, arrange them in time order to establish a growth trend, and generate a length growth sequence.

[0024] S302: Based on the length growth sequence, obtain the crack width data within the corresponding period, perform the difference operation between adjacent periods, construct a width growth record consistent with the length sequence time, and perform first-order growth rate calculation within the continuous period for the two sets of sequences respectively to obtain the crack growth rate set.

[0025] S303: For the set of crack growth rates, compare the growth directions of length and width over time to see if they are consistent, and determine if the difference in growth rate is continuously decreasing. Filter out the periodic segments that meet the conditions of synchronous increase and decreasing difference, locate the corresponding expansion behavior, and generate a crack synchronous expansion stage identifier.

[0026] As a further aspect of the present invention, the specific steps of S4 are as follows:

[0027] S401: Call the time segments recorded in the crack initiation region signal, the crack outline expansion path and the crack synchronous expansion stage identifier, extract the corresponding timestamp set, perform the intersection operation on the three sets in chronological order, filter all time segments that appear in at least two items in a continuous period, and generate a cross-active time segment group.

[0028] S402: Based on the cross-active time segment group, determine whether the number of corresponding result types in each time segment is not less than two, set and check whether the time period is at the active threshold in all results, mark the time interval that meets the common active characteristics of multiple results, and obtain the set of active intervals of multiple indicators.

[0029] S403: Based on the set of active intervals of the multiple indicators, verify the periodic segments in which active results exist continuously in all time segments, assign evolutionary status identifiers to the marked segments, merge and organize them into time-series data, and generate records of active segments of crack evolution.

[0030] As a further aspect of the present invention, the specific steps of S5 are as follows:

[0031] S501: Call the crack displacement change, boundary offset direction sequence and width growth rate sequence corresponding to the active time period contained in the crack evolution active segment record, group the crack behaviors with continuous trends into the same group according to the monitoring area and time order, divide the groups according to the time series number, and generate crack behavior grouping index.

[0032] S502: Based on the crack behavior grouping index, determine whether the offset direction of the crack remains consistent within the period, compare the width change trend within each group, filter the behavior set that simultaneously satisfies the consistency of direction and the continuous increase of width, and obtain the crack trend consistency marker group.

[0033] S503: For the crack trend consistency marker group, determine that those that meet the trend conditions are classified as extended class, and the rest are classified as stable class. Summarize the category markers and time information of multiple groups, construct a structured record table with time period index and classification marker, and generate crack evolution classification label set.

[0034] A full life-cycle monitoring system for cracks in concrete structures, including:

[0035] The acoustic stress monitoring module is used to achieve S1: deploy acoustic points at the downstream dam pier nodes, measure the propagation time to establish a time series, calculate the rate of change of sound velocity, obtain the concrete stress change value, compare the stress curve with the fracture reference curve, and generate the crack initiation area signal.

[0036] The crack contour recognition module is used to implement S2: acquire the inner wall image sequence of the crack initiation region signal, extract boundary feature points, calculate the Euclidean distance and angle change of the differential time, and if the direction is continuous and the distance increases, generate the crack contour expansion path.

[0037] The crack synchronous propagation detection module is used to implement S3: Based on the crack profile propagation path, extract the crack length and width sequence of the main span bottom edge crack, calculate the growth rate, determine whether the two are rising synchronously and the difference is decreasing, and generate a crack synchronous propagation stage identifier.

[0038] The crack evolution active cycle extraction module is used to implement S4: call the crack initiation region signal, the crack outline expansion path and the crack synchronous expansion stage identifier, perform intersection operation according to the timestamps of the three sets of results, filter the cycles that meet the active threshold, and generate crack evolution active segment records;

[0039] The crack evolution classification module is used to implement S5: based on the records of the active crack evolution segments, analyze the changes in crack displacement, direction and width in time series, classify those with consistent direction and continuously increasing width into the extended class, and the rest into the stable class, and generate a crack evolution classification label set.

[0040] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0041] In this invention, stress evolution is reflected by acoustic time difference, crack propagation trend is extracted by combining image feature point trajectory, active state is identified by superimposing crack size growth rate synchronization, high-frequency change segments are located by multi-index time intersection, and crack behavior type is classified according to displacement direction and width growth trend. This enables continuous tracking and classification labeling of the entire process from crack initiation to evolution, enhances the response capability to abnormal crack behavior, and improves the accuracy and timeliness of structural safety status identification. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a schematic diagram of the steps of the present invention;

[0044] Figure 2 This is a detailed schematic diagram of S1 of the present invention;

[0045] Figure 3 This is a detailed schematic diagram of S2 of the present invention;

[0046] Figure 4 This is a detailed schematic diagram of S3 of the present invention;

[0047] Figure 5 This is a detailed schematic diagram of S4 of the present invention;

[0048] Figure 6 This is a detailed schematic diagram of S5 of the present invention;

[0049] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0050] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0051] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0052] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0053] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0054] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0055] Please see Figure 1 This invention provides a method for monitoring cracks in concrete structures throughout their entire life cycle, comprising the following steps:

[0056] S1: Sound wave transmitting and receiving points are set up at the downstream pier nodes of the concrete dam. A time series is established by measuring the sound wave propagation time between nodes. The sound velocity change rate between node pairs is calculated according to the time series. The load value of the stress state of the concrete structure at the corresponding node is obtained as a function of time. Time difference processing is performed to generate stress change curves. The difference between sound velocity change and stress change is compared one by one at each time. The fracture monitoring benchmark curve is used as the comparison reference threshold to determine whether the upper limit of change critical value has been reached and to generate crack initiation area signal.

[0057] S2: Obtain differentiated time period image sequences on the inner wall surface of the tunnel covered by the signal in the crack initiation area, extract the curvature reversal points and abrupt change direction points in the boundary contour curve to construct a feature point set, calculate the Euclidean distance and angle change value between corresponding feature points in adjacent time periods, construct a set of offset direction trajectories arranged in time order, determine whether the offset direction is consistent in continuous time periods and accompanied by a distance increasing trend, if the trend continues and the path is stable, generate the crack contour expansion path;

[0058] S3: Call the bridge main span bottom edge area covered by the crack outline expansion path, obtain the crack length and width values ​​in the current period, perform time difference calculation to generate length growth sequence and width growth sequence respectively, perform one growth rate calculation on the above sequences in a continuous period, compare whether the direction of numerical change is consistent, and determine whether the growth rate difference between the two sets of sequences in a continuous period continues to decrease. If the two sets of sequences move upward synchronously and the difference decreases, generate a crack synchronous expansion stage identifier.

[0059] S4: Call the time segments recorded in the crack initiation region signal, crack outline propagation path and crack synchronous propagation stage identifier, perform intersection operation according to the timestamps of the three sets of results, filter the time period in the continuous cycle where at least two results are active at the same time, determine whether the time period is a multi-indicator cross-active interval, if the match is true, lock the time segment as a high-frequency change segment, and generate crack evolution active segment record.

[0060] S5: Call the crack displacement change value, boundary offset direction and width growth rate corresponding to the active time period in the crack evolution active segment record, classify the crack behavior with continuous trend characteristics in the same monitoring area into a group according to the time series, determine whether the crack growth direction is consistent in multiple periods and whether the width change continues to grow upward. If the direction and width trends are stable at the same time, it is classified into the crack expansion class; otherwise, it is classified into the stable class, and a crack evolution classification label set is generated.

[0061] The crack initiation region signal includes the upper limit threshold of sound velocity change, the deviation of stress change, and the critical value triggering judgment result. The crack profile expansion path includes the feature point offset direction, the distance change result between feature points, and the path stability judgment result. The crack synchronous expansion stage identifier includes the crack length growth rate, the crack width growth rate, and the trend of the growth rate difference. The crack evolution active segment record includes the multi-index cross time period, the crack state activity level, and the high-frequency evolution time segment. The crack evolution classification annotation set includes the crack expansion direction, the crack width growth trend, and the behavioral continuity characteristics.

[0062] Please see Figure 2 The specific steps of S1 are as follows:

[0063] S101: Set up sound wave transmitting and receiving points at the downstream pier nodes of the concrete dam, obtain the start and arrival times of sound wave propagation between nodes, establish node pairs based on the relationship between propagation duration and distance between nodes, and generate a sound wave propagation time series group.

[0064] First, the geometric location of the piers needs to be clearly defined. Based on the dam structural drawings and construction measurement data, the layout of each pier node is determined. Then, an acoustic sensor is installed at each node. The sensor consists of two parts: an acoustic transmitter and a receiver. The transmitter emits acoustic pulses of a specific frequency in a directional manner. The frequency is usually set to around 40kHz based on the wave velocity range inside the concrete. The receiver needs to have its time synchronization function enabled to ensure accurate time stamps for each transmission and reception. After the sensors are deployed, the acoustic transmission experiment for each pair of nodes is triggered by the control module. In each experiment, the transmitter automatically records the transmission time when it starts the acoustic signal, and the receiver records the reception time when it receives the acoustic wave. The time difference between the two is the propagation time. The three-dimensional spatial coordinates should be obtained in advance, for example, using a laser rangefinder or a 3D modeling scanner. These coordinates are used to calculate the distance between node pairs. After determining the spatial distance between the transmitting and receiving points, the node pair data records are formed by combining them with the propagation time. After deploying multiple node pairs, each pair needs to be activated in sequence for multiple measurement operations. The standard setting is to measure each node pair three times, and after filtering out outliers, the average value is taken. The time difference between the three recorded values ​​must not exceed the set error limit, which is initially set to 0.005 milliseconds. If it exceeds the limit, it needs to be measured again. Finally, the propagation time and spatial distance of all node pairs are summarized into a complete propagation time series group. All record numbers are arranged in the measurement order to form a sound wave propagation time series group.

[0065] S102: Based on the sound wave propagation time series group, call the node pairs to change the propagation time, calculate the ratio of propagation time per unit time to node distance, aggregate the results of the region, and obtain the sound speed change rate distribution map;

[0066] For each pair of nodes, the difference in propagation time at consecutive moments is calculated. This is achieved by dividing the change in propagation time between two adjacent time points by their corresponding time intervals, thus obtaining the rate of change of propagation time per unit time. Given the distance between each pair of nodes, the ratio of propagation time to distance can be further calculated to obtain a relative index of sound speed change. In practice, multiple calculations are performed on each pair of nodes to obtain the rate of sound speed change over different time periods. These results are then aggregated according to the region to which the node pair belongs. Region division is based on the structural location of the nodes, typically using structural component zoning, stress concentration zones, and crack-prone areas as reference standards. During aggregation, each data set needs to be assigned a different weight, with the weight setting based on the region. The stress sensitivity level of the domain structure is initially set as follows: the weight of areas with significant stress response is 1.5, the weight of ordinary areas is 1.0, and the weight of areas with weak structural constraints is 0.8. The sound velocity change rate index of each area is obtained by weighted averaging. Multiple areas are compared to form a distribution map. The map is arranged horizontally according to the node area number, and the vertical axis corresponds to the magnitude of the sound velocity change ratio of each area. Three standards are set in the interval determination: the stable area is set as a change rate of less than 0.05 milliseconds per meter, the change area is 0.05 to 0.1 milliseconds per meter, and the abnormal area is greater than 0.1 milliseconds per meter. This interval is determined based on measured engineering experience and the analysis results of historical crack precursors, and serves as an important reference standard for identifying structural state fluctuations. The final output is a sound velocity change rate distribution map.

[0067] S103: Based on the sound velocity change rate distribution map and the structural response coefficient of the nodal region, calculate the force change per unit area according to the time series, analyze the force function of the nodal region, and generate the crack initiation region signal.

[0068] To further calculate the change in stress per unit area for each region, firstly, for each node region, the corresponding structural response coefficient is found based on its sound velocity change rate index. This coefficient is set according to the concrete grade, component type, stress direction, and reinforcement arrangement. For the foundation, the coefficient is set as follows: C30 strength grade corresponds to a coefficient of 1.0, C40 corresponds to 1.2, densely reinforced areas can be increased by 0.1 to 0.2, and sparsely reinforced areas or non-principal stress directions can be decreased to 0.8. This coefficient can be calibrated through structural tests or finite element simulation results. After determining this coefficient, the sound velocity change rate value of that region is multiplied by the structural response coefficient to obtain the change in stress per unit area for that region. Then, the stress change values ​​at different time nodes are plotted as a continuous curve in time sequence. Finally, based on the slope of the stress curve, i.e. The rate of stress increase per unit time is used to determine the stress growth trend. For each time period, the difference in stress value between adjacent time points is calculated and divided by the time difference to obtain the stress change rate. When this rate is continuously higher than a set threshold and the duration exceeds the specified time period, it can be determined that there is an abnormal stress concentration trend in the area. The determination of crack initiation signal is based on whether this rate of change is greater than the crack sensitivity threshold. The sensitivity threshold is set to 0.1 Newtons per square meter per second based on a large amount of monitoring data before the formation of cracks in concrete structures. The duration threshold is initially set to 15 minutes. If the stress change rate in a certain area is 0.12 and lasts for 20 minutes, the judgment condition is met, and the area can be marked as the crack signal initiation area. Finally, the area number, spatial location and maximum stress rate are recorded to form a set of crack initiation area signals.

[0069] Please see Figure 3 The specific steps of S2 are as follows:

[0070] S201: Obtain differentiated time period image sequences on the inner wall surface of the tunnel covered by the signal in the crack initiation area, extract all curvature reversal points and abrupt change direction points of the boundary contour curve in each image, combine them to form a set of feature point coordinates for the corresponding time period, and generate a time period feature point set.

[0071] Based on the crack early warning signal triggered by monitoring, the image acquisition time is divided into multiple continuous monitoring periods, for example, images are acquired every 8 hours. During the acquisition process, it is ensured that the same location, shooting angle, and lighting conditions are consistent. High-resolution industrial cameras are used to acquire images of the structural surface. The acquired images are first processed for noise reduction and contrast enhancement to improve the accuracy of subsequent boundary extraction. In each image, the edge contour of the tunnel inner wall is extracted. After edge extraction, the boundary contour curve needs to be traversed. On each segment of the curve, the curvature direction change trend is continuously judged. When two consecutive curvature values ​​show a sign change, and the angle formed by three adjacent points exceeds the set minimum reversal threshold, the point is marked as a curvature reversal point. The initial reversal threshold is set to 12 degrees, derived from the statistical results of the curvature change of the minimum crack edge inflection point in the experiment. The point is further judged to determine whether its location is in a sudden change corner area. If three adjacent points form a sudden change in angle and the change value exceeds the set sudden change angle threshold of 25 degrees, it is marked as a sudden change direction point. This threshold is obtained by statistical analysis of the sudden change direction identified by manually annotated real crack images. It is a typical inflection point feature in the process of crack contour evolution. All feature points extracted according to the above rules need to record their positions in the image coordinate system. Then, they are converted into actual spatial coordinates by combining the image projection transformation relationship. For example, if the image resolution is 0.25 mm per pixel, then the spatial position of pixel (1200, 850) after conversion is (300 mm, 212.5 mm). This operation is repeated for all images to extract and combine the feature point coordinates of the current time period. Each image generates an independent feature point set, and finally forms multiple time period feature point sets arranged in chronological order.

[0072] S202: Based on the set of feature points in a time period, calculate the Euclidean distance and the change in the angular direction between corresponding feature points in adjacent time periods, arrange them in order to construct the temporal offset path of the feature points, and compare the changes in direction and distance of the path over time to obtain a continuous offset direction sequence;

[0073] A pairing operation is performed on the feature point sets of two adjacent time periods. The pairing rule is to establish a correspondence between two points whose spatial distance is within a preset error range. This error range is initially set to 20 mm, and this value is derived from the maximum static feature point offset value observed in the early stage of actual tunnel crack development. During the pairing process, based on the coordinates of each feature point in the current time period, all points in the next time period are traversed, and the nearest point that meets the error range is found as the pairing object. After establishing the starting point pair, the actual spatial distance and the change in the direction of the connecting line between corresponding feature points in the two time periods are recorded respectively. The angular change of the direction of the connecting line between adjacent feature points is judged by a three-point vector. The direction of the connecting line in the previous time period is compared with the direction of the connecting line in the current time period. If the angle between the directions is greater than 5 degrees, it is considered that there is a change in direction. The included angle value is recorded as the offset direction angle. After all feature points are paired and calculated continuously, the paired data are arranged into an offset path sequence in chronological order. Each path segment records its start and end time, start and end point positions, displacement distance, and direction angle information. For example, a feature point is located at (1.25m, 0.85m) in time period t1 and moves to (1.29m, 0.87m) in time period t2, with a distance of 5.0 mm and a direction change of 16 degrees. This segment constitutes a path segment. All path segments constitute a complete feature point offset path. All segments in the path are arranged in chronological order, and the direction change trend and distance change magnitude of each path segment are further statistically analyzed. The offset information of each segment is extracted in sequence and organized into a continuous offset direction sequence, which serves as the basis for subsequent crack profile trend determination.

[0074] S203: Based on the continuous offset direction sequence, determine whether the path direction is continuous and consistent and whether the corresponding distance has an increasing trend, filter the path segments that meet the stable offset threshold, locate the contour boundary change trend of the offset concentration area, and generate the crack contour expansion path.

[0075] For each offset path segment, a trend consistency judgment is performed. The judgment process first involves statistically analyzing the continuous changes in the direction of the direction change angle within the path. If three or more consecutive segments have the same sign for the direction change angle, and the angle change value fluctuates within a range of less than 20 degrees, the direction is considered continuous and consistent. This range is derived from on-site artificial crack evolution annotation samples and serves as the benchmark value for judging the stability of the crack development path. Subsequently, the distance changes between adjacent segments in the path are compared to determine if there is an increasing trend. The specific criterion is that the displacement distance of the later segment minus the previous segment must be greater than 3 mm. If three consecutive segments meet this condition, the path segment is judged to have a stable increasing trend. The 3 mm threshold is set based on the statistical analysis of the daily average displacement velocity changes of concrete cracks in the early stages of expansion. If the path segment meets both the conditions of continuous direction and increasing distance, it is selected as a stable offset path segment. Then, the start and end points of all selected path segments are connected and marked as possible crack propagation paths. By analyzing the boundary change characteristics in the image where the path segment is located, the evolution trajectory of the contour boundary over multiple time periods is further drawn. For example, a stable path starts at the feature point (1.10m, 0.75m) in image number t3 and ends at the feature point (1.26m, 0.84m) in image number t6. The corresponding path consists of three segments with direction change values ​​of 14 degrees, 12 degrees, and 15 degrees, and displacement values ​​of 4.1 mm, 6.9 mm, and 10.3 mm, respectively, all of which meet the judgment criteria. Finally, this path is marked as a crack contour propagation path.

[0076] Please see Figure 4 The specific steps of S3 are as follows:

[0077] S301: Call the bridge main span bottom edge area covered by the crack outline expansion path, obtain the crack length and width data in the current period area, perform difference calculation on the crack length data of two adjacent periods, arrange them in time order to establish the growth trend, and generate a length growth sequence.

[0078] First, the crack path is spatially matched in the 3D structural model based on its spatial coordinate range, placing the path at a specific location on the structure. For example, the path corresponds to the area within the bridge's mid-span bottom slab, from 1.2 meters to 3.6 meters on the X-axis and within ±0.5 meters of the bridge's longitudinal centerline on the Y-axis. Crack parameters collected by all crack monitoring devices or visual monitoring within this area must be used as the input data source for the current analysis cycle. Each cycle's data must record core information such as crack number, measurement time, crack length, and width. After recording the length of each crack, a difference calculation is performed between adjacent cycles. That is, the crack length of the later cycle minus the length of the previous cycle forms the crack growth. If the crack length L03 in cycle T1 is 154 mm and the measurement in cycle T2 is 161 mm, then the length growth for that cycle is... The length is 7 mm. All crack data are calculated in this way, and the length growth sequence is formed by sorting them in the order of monitoring time. The recording format is {T1→T2: L03: +7mm, L05: +3mm; T2→T3: L03: +5mm, L05: +2mm}. It should be noted that if a crack has a growth value of 0 or negative in a continuous period, it should be marked as no growth or closed and excluded from the growth sequence to ensure the accuracy of trend judgment. The rule for judging outliers in the data is that the growth value is more than twice the historical average, or it is negative growth for more than two periods. Such data will be archived separately and will not participate in the trend analysis operation in this stage. The final length growth sequence will be used as the benchmark sequence input for subsequent width growth and growth rate calculations, fully reflecting the current period's crack length evolution trend.

[0079] S302: Based on the length growth sequence, obtain the crack width data within the corresponding period, perform the difference operation between adjacent periods, construct a width growth record consistent with the length sequence time, and perform first-order growth rate calculation within the continuous period for the two sets of sequences to obtain the crack growth rate set.

[0080] Extract the crack width data recorded in each cycle. The width data can be obtained from manual measurement, inductive crack gauge, or image recognition. Each width value must point to a specific crack number and timestamp. Calculate the difference between adjacent cycles for the extracted width data and record it as the crack width growth value within that cycle. For example, if the width of crack L03 is 0.38 mm in cycle T1 and 0.43 mm in cycle T2, the growth value is 0.05 mm. The width growth record maintains the same chronological order as the length growth and is constructed as a width growth record sequence, such as {T1→T2: L03: +0.05 mm; T2→T3: L03: +0.03 mm}. After completing the width data difference recording, calculate the first-order growth rate for both the length and width growth sequences during consecutive cycles. Specifically, subtract the growth rate of the previous cycle from the current cycle's growth rate to reflect the increase. The rate of change over a long period, for example, if the length increases by +7mm from T1 to T2 and by +5mm from T2 to T3, then the first-order growth rate is -2mm. If the width increases by +0.05mm and +0.03mm, then the first-order growth rate is -0.02mm. These growth rate data are compiled to form a crack growth rate set. Each set records the corresponding period, crack number, length growth rate value, and width growth rate value. To ensure the validity of the data, a data filtering range needs to be set. Any record with an absolute growth rate value less than 0.01mm is considered a record of insufficient change and is marked to avoid interference from minor changes in the judgment. This threshold is set based on twice the sensor measurement accuracy and image recognition tolerance, that is, corrected upwards from a measurement accuracy of 0.005mm. Finally, the crack growth rate set in time order is obtained, which facilitates subsequent trend consistency and synchronous expansion identification.

[0081] S303: For the set of crack growth rates, compare the growth directions of length and width over time to see if they are consistent, and determine if the difference in growth rates is continuously decreasing. Filter out the periodic segments that meet the conditions of synchronous increase and decreasing difference, locate the corresponding expansion behavior, and generate a crack synchronous expansion stage identifier.

[0082] The growth rates of crack length and width are compared and analyzed periodically in chronological order. First, it is determined whether their growth directions are consistent, i.e., whether their numerical signs are the same. If the growth rates of both length and width are positive or both are negative within a certain period, they are considered to be in the same direction. For example, in the T2→T3 period, the growth rate of crack L03's length is -2mm and its width is -0.02mm, both negative, satisfying the consistency judgment condition. Then, the difference between the growth rates of length and width in that period is calculated, i.e., the difference between their absolute values, and it is determined whether this difference shows a decreasing trend in consecutive periods. The judgment condition for a decreasing difference in consecutive periods is that the absolute difference between the differences of two adjacent periods is greater than 0.00. The threshold of 5mm is the average standard value of the actual measurement error, set based on the smallest distinguishable expansion feature in the historical data of cracks in 20 bridges. For example, the difference in the growth rate of the length and width of crack L03 in the T2→T3 cycle is 1.98mm, and the difference in the T3→T4 cycle is 1.72mm. The difference between the two cycles is reduced to 0.26mm, which meets the judgment of the shrinking trend. If the two conditions of consistent direction and shrinking difference are met for three consecutive cycles, the cycle segment is marked as a synchronous expansion segment. All cycle segments that meet the screening conditions are numbered, marked and their start and end times and corresponding crack numbers are recorded. Finally, the synchronous expansion stage identifier of crack is output, which is used to further locate the time period in the structure where significant deformation development may occur.

[0083] Please see Figure 5 The specific steps of S4 are as follows:

[0084] S401: Call the time segments recorded in the crack initiation region signal, crack profile propagation path and crack synchronous propagation stage identifier, extract the corresponding timestamp set, perform the intersection operation on the three sets in chronological order, filter all time segments that appear in at least two items in a continuous period, and generate a cross-active time segment group.

[0085] First, the corresponding timestamp sets need to be extracted from the three types of records. Specifically, the signal trigger time and duration period need to be extracted from the crack initiation signal; the evolution time period corresponding to the path generation needs to be extracted from the contour expansion path record; and the start and end periods of the synchronization trend need to be extracted from the synchronization expansion identifier. After unifying the format of the three types of data, they are all standardized into structured time segment groups with start time, end time, crack number, and data source label. Then, the three sets are converted into time series sets with time period number as the primary key, and numbered and sorted in chronological order. The intersection operation is performed on the three sets on a unified time axis. The operation method is to compare by period number and count the number of times each period number appears in the three sets. If a period number appears in the three sets, the intersection operation is performed. If a period number appears less than twice, it is included in the candidate cross-period list. The data in the candidate period list is judged for continuity. The judgment criteria are whether the period numbers are consecutive and the number is not less than two periods. For example, if the period numbers T6, T7, and T8 meet the above conditions, they are recorded as a cross-active time segment group. This group is defined as a time period in which multiple features overlap during the crack evolution stage. The standard of not less than two consecutive periods is determined by combining the minimum stable crack evolution period in the project. Based on the time span from the onset of the characteristic crack to the formation of a stable expansion trend in the long-term monitoring data of multiple bridges, the shortest is 48 hours, which is two monitoring periods. Therefore, this is used as the screening threshold standard. Finally, all cross-active time segment groups that meet the screening conditions are output.

[0086] S402: Based on the cross-active time segment group, determine whether the number of corresponding result types in each time segment is not less than two, set and check whether the time period is at the active threshold in all results, mark the time interval that meets the common active characteristics of multiple results, and obtain the set of active intervals of multiple indicators.

[0087] For each time period, the number of types and the validity of the status are checked. The first step is to determine whether the number of result types corresponding to that time period is not less than two. That is, the three types of results corresponding to that time period are compared one by one to see if they are recorded in any two of the crack initiation signal, profile propagation path and synchronous propagation stage. For example, if the T8 cycle appears in the crack initiation signal and profile propagation path, then the number of types is 2, which meets the basic requirements. Next, it is checked whether the time period is marked as "active" in the relevant records. The active status is determined by analyzing the recorded data values. If the recorded values ​​are parameters such as displacement growth, crack length or width growth, structural strain and If the change exceeds a preset threshold, it is considered an active state. The displacement growth threshold is set to 2 mm, the length growth threshold to 5 mm, and the width growth threshold to 0.02 mm. These thresholds are based on the minimum significant change of each monitoring parameter, and their values ​​are derived from statistical analysis of the annual average structural response deformation monitoring data of 30 bridges. If at least two types of results are confirmed to occur within a time period and both are active, that time period is marked as a multi-indicator active interval. All time segments that meet the conditions will be uniformly summarized and numbered, for example, from MIAQ_01 to MIAQ_n, forming a set of multi-indicator active intervals for unified use in subsequent crack evolution state identification.

[0088] S403: Based on the set of active intervals of multiple indicators, verify the periodic segments in which active results exist continuously in all time segments, assign evolution status labels to the marked segments, merge and organize them into time-series data, and generate records of active segments of crack evolution.

[0089] First, the start times of all active interval numbers are sorted by time to determine if there are any periodic interruptions. If there are no missing periods between the end time of two interval numbers and the start time of the next interval, it is considered a continuous active segment; otherwise, it is split into two independent active segments. Then, all periods in each continuous active segment are checked to determine if they are in a valid "activated" state in the original three types of results. The activation state is verified by checking whether the change in crack length or width in the original record meets the condition of continuously exceeding the above-mentioned set threshold. For example, if the length change is greater than 5 mm for three consecutive periods or the width change is greater than 0.03 mm for three consecutive periods, it is confirmed that it is continuously in the evolutionary state. If any indicator in a period is lower than the threshold, it is considered to be in an evolutionary state. If a threshold is not met and the occurrences are not continuous, the cycle is removed from the continuous segment to maintain the purity of the segment. In the confirmed continuous cycle segments, each segment is assigned an evolution status identifier. The status classification criteria are as follows: "accelerated evolution" is defined as a cycle number ≥ 5 with an average growth value higher than 150% of the historical average; "stable evolution" is defined as a cycle number 3-4 with a growth value fluctuation less than 50% of the historical average; and "initial evolution" is defined as a cycle number ≤ 2. This classification criterion is based on the analysis of long-term crack development trends in structural monitoring and the standard values ​​set for bridge inspection. Finally, all continuous time segments with evolution status identifiers are organized into a structured record data table, which includes: number, start and end cycle number, crack number, activation source, evolution type, and forms a complete record of active crack evolution segments.

[0090] Please see Figure 6 The specific steps of S5 are as follows:

[0091] S501: Call the crack displacement change, boundary offset direction sequence and width growth rate sequence corresponding to the active time period contained in the crack evolution active segment record, group the crack behaviors with continuous trends into the same group according to the monitoring area and time sequence, divide the groups according to the time series number, and generate crack behavior grouping index.

[0092] First, the three types of data sources need to be uniformly formatted. The period number, displacement data, offset angle sequence, and width change value of each crack are organized into a structured dataset in chronological order. Crack displacement changes are expressed in millimeters, recording the change in the crack center point coordinates between two adjacent monitoring periods. The crack boundary offset direction sequence is recorded as polar coordinate angles, representing the direction of crack boundary movement. The width increase is the difference in crack opening values ​​between two adjacent periods, expressed in millimeters. For example, crack number C07 has a width of 0.38 mm in period T3 and 0.43 mm in period T4, representing an increase of 0.05 mm. Then, all crack data are categorized by structural region. The region numbering is standardized according to bridge structure classification standards, such as the main span... The segment is designated as R01, and the main beam connection segment is designated as R02. Within each region, cracks are sorted by period number from smallest to largest. It is checked whether there are non-zero displacements, continuous and non-reversed direction angle sequences, and positive width growth values ​​within three or more consecutive periods. If these conditions are met, it is determined to have a continuous trend. This condition is based on the common stage evolution behavior in the development of cracks in actual bridges. Usually, a stable trend must be observed within 72 consecutive hours to be considered categorizable. Therefore, the minimum number of consecutive periods is set to three. After this condition is met, the crack record is assigned to the same group and marked with a group number, such as behavior group BG01. The record content includes: crack number, start period, end period, region number, and the corresponding three types of data sequences. All grouping information is uniformly organized into a crack behavior grouping index.

[0093] S502: Based on the crack behavior grouping index, determine whether the offset direction of the crack remains consistent within the period, compare the width change trend within each group, filter the behavior set that simultaneously satisfies both direction consistency and continuous width growth, and obtain the crack trend consistency marker group.

[0094] First, assess whether the offset direction remains consistent within the cycle. The standard for consistent direction is that the change in offset angle between cycles must not exceed a set angle threshold, and the direction of angle change must remain unidirectional. For example, if the crack offset angle within a group is 142° in T3, 145° in T4, and 149° in T5, then the direction is continuously clockwise, and the angle change is less than 10 degrees, meeting the consistency standard. This angle threshold is set at ±15 degrees, determined based on the minimum difference in the actual offset change of the crack boundary in the field monitoring data. The purpose of this setting is to exclude slight directional fluctuations caused by occasional non-structural changes. Next, determine the width change trend within the group. If the width changes continuously within the group... If the crack width growth value is always positive and the increment is not less than the minimum identification threshold of 0.02 mm, it is determined that the width is continuously increasing. This threshold is derived from the minimum measurable increment of the sensor and image recognition in the early stage of dynamic crack propagation in multiple concrete structure monitoring experiments. Cracks that meet this condition are classified into the trend consistency label group. The records after grouping must completely include the original behavior group number, crack number, monitoring period range, data on the change of offset direction in each period and width growth record. All trend consistency crack behavior sets that meet the judgment conditions are identified by a unified numbering method, such as TCMG_001, TCMG_002, etc., and the crack trend consistency label group is output.

[0095] S503: For crack trend consistency marker groups, determine that those that meet the trend conditions are classified into the extended class, and the rest are classified into the stable class. Unify and summarize the category markers and time information of multiple groups, construct a structured record table with time period index and classification marker, and generate a crack evolution classification label set.

[0096] Further trend classification judgment is performed on each group of behaviors. Specifically, it is checked whether it simultaneously meets the two indicators of directional consistency and continuous width growth. If both are met, the group is marked as an expansion class; otherwise, it is marked as a stable class. During the judgment, if the offset angle of any cycle reverses more than the set threshold, or if the width growth value of any cycle is zero or negative, then even if other cycles meet the growth trend, the group is still marked as a stable class. This classification standard aims to clearly distinguish between the evolving and temporarily stable crack states in the structure. After the classification is completed, the classification results of all crack behavior groups are uniformly organized into a structural record table. The record fields include: crack number, group number, region number, start cycle number, end cycle number, maximum displacement increment, total width growth, maximum offset angle change, and classification label. The classification label is set to only two types: "expansion class" or "stable class". Finally, all records are sorted in a secondary order according to time and region number to form a complete temporal structure table and generate a crack evolution classification label set for further hierarchical processing and regional evolution tracking in monitoring.

[0097] Please see Figure 7A full life-cycle monitoring system for cracks in concrete structures, including:

[0098] The acoustic stress monitoring module is used to achieve S1: deploy acoustic points at the downstream dam pier nodes, measure the propagation time to establish a time series, calculate the rate of change of sound velocity, obtain the concrete stress change value, compare the stress curve with the fracture reference curve, and generate the crack initiation area signal.

[0099] The crack contour recognition module is used to achieve S2: acquire the inner wall image sequence of the crack initiation region signal, extract boundary feature points, calculate the Euclidean distance and angle change of the differential time, and generate the crack contour extension path if the direction is continuous and the distance increases.

[0100] The crack synchronous propagation detection module is used to implement S3: based on the crack profile propagation path, extract the crack length and width sequence of the bottom edge of the main span, calculate the growth rate, determine whether the two are rising synchronously and the difference is decreasing, and generate a crack synchronous propagation stage identifier.

[0101] The crack evolution active cycle extraction module is used to implement S4: call the crack initiation region signal, the crack outline expansion path and the crack synchronous expansion stage identifier, perform intersection operation according to the timestamps of the three sets of results, filter the cycles that meet the active threshold, and generate crack evolution active segment records;

[0102] The crack evolution classification module is used to implement S5: based on the records of active crack evolution segments, it analyzes the changes in crack displacement, direction and width in time series, classifies cracks with consistent direction and continuously increasing width into the extended class, and the rest into the stable class, generating a crack evolution classification label set.

[0103] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for monitoring cracks in concrete structures throughout their entire life cycle, characterized in that, Includes the following steps: S1: Set up acoustic points at the downstream pier nodes of the dam, measure the propagation time to establish a time series, calculate the rate of change of sound velocity, obtain the value of concrete stress change, compare the stress curve with the fracture reference curve, and generate the crack initiation area signal. S2: Obtain the inner wall image sequence of the crack initiation region signal, extract boundary feature points, calculate the Euclidean distance and angle change of the differential time, and if the direction is continuous and the distance increases, generate the crack outline expansion path. S3: Based on the crack profile expansion path, extract the crack length and width sequence at the bottom edge of the main span, calculate the growth rate, determine whether the two increase synchronously and the difference decreases, and generate a crack synchronous expansion stage identifier. S4: Call the crack initiation region signal, the crack outline expansion path and the crack synchronous expansion stage identifier, perform intersection operation based on the timestamps of the three sets of results, filter the cycles that meet the active threshold, and generate crack evolution active segment records; S5: Based on the records of active crack evolution segments, analyze the changes in crack displacement, direction and width in time series, classify those with consistent direction and continuously increasing width into the extended class, and the rest into the stable class, and generate a crack evolution classification label set.

2. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, The crack initiation region signal includes the upper limit threshold of sound velocity change, the deviation amplitude of stress change, and the critical value triggering judgment result. The crack profile expansion path includes the feature point offset direction, the distance change result between feature points, and the path stability judgment result. The crack synchronous expansion stage identifier includes the crack length growth rate, the crack width growth rate, and the trend of the growth rate difference. The crack evolution active segment record includes the multi-index cross time period, the crack state activity level, and the high-frequency evolution time segment. The crack evolution classification annotation set includes the crack expansion direction, the crack width growth trend, and the behavioral continuity characteristics.

3. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, The definition of synchronous increase and decreasing difference is that the length and width of the crack at the bottom edge of the main span increase together over time, and the difference between their growth gradually decreases.

4. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, If the crack direction is consistent with the definition of the extended class and the stable class and the width continues to increase, it is determined to be an extended class; otherwise, it is a stable class.

5. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Set up sound wave transmitting and receiving points at the downstream pier nodes of the concrete dam, obtain the start and arrival times of sound wave propagation between nodes, establish node pairs based on the relationship between propagation duration and distance between nodes, and generate a sound wave propagation time series group. S102: Based on the sound wave propagation time series group, call the node pairs to change the propagation time, calculate the ratio of the propagation time per unit time to the node distance, aggregate the results of the region, and obtain the sound speed change rate distribution map; S103: Based on the sound velocity change rate distribution map and the structural response coefficient of the nodal region, calculate the force change per unit area according to the time series, analyze the force function of the nodal region, and generate the crack initiation region signal.

6. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Obtain a sequence of differentiated time period images on the inner wall surface of the tunnel covered by the signal in the crack initiation area, extract all curvature reversal points and abrupt change direction points of the boundary contour curve in each image, combine them to form a set of feature point coordinates for the corresponding time period, and generate a set of feature points for the time period. S202: Based on the set of feature points in the time period, calculate the Euclidean distance and the change in the included angle between corresponding feature points in adjacent time periods, arrange them sequentially to construct the temporal offset path of the feature points, and compare the changes in direction and distance of the path according to time to obtain a continuous offset direction sequence; S203: Based on the continuous offset direction sequence, determine whether the path direction is continuous and consistent and whether the corresponding distance has an increasing trend, filter the path segments that meet the stable offset threshold, locate the contour boundary change trend of the offset concentration area, and generate the crack contour expansion path.

7. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Call the bridge main span bottom edge area covered by the crack outline expansion path, obtain the crack length data and width data in the current period area, perform difference calculation on the crack length data of two adjacent periods, arrange them in time order to establish a growth trend, and generate a length growth sequence. S302: Based on the length growth sequence, obtain the crack width data within the corresponding period, perform the difference operation between adjacent periods, construct a width growth record consistent with the length sequence time, and perform first-order growth rate calculation within the continuous period for the two sets of sequences respectively to obtain the crack growth rate set. S303: For the set of crack growth rates, compare the growth directions of length and width over time to see if they are consistent, and determine if the difference in growth rate is continuously decreasing. Filter out the periodic segments that meet the conditions of synchronous increase and decreasing difference, locate the corresponding expansion behavior, and generate a crack synchronous expansion stage identifier.

8. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Call the time segments recorded in the crack initiation region signal, the crack outline expansion path and the crack synchronous expansion stage identifier, extract the corresponding timestamp set, perform the intersection operation on the three sets in chronological order, filter all time segments that appear in at least two items in a continuous period, and generate a cross-active time segment group. S402: Based on the cross-active time segment group, determine whether the number of corresponding result types in each time segment is not less than two, set and check whether the time period is at the active threshold in all results, mark the time interval that meets the active threshold, and obtain the set of active intervals of multiple indicators. S403: Based on the set of active intervals of the multiple indicators, verify the periodic segments in which active results exist continuously in all time segments, assign evolutionary status identifiers to the marked segments, merge and organize them into time-series data, and generate records of active segments of crack evolution.

9. The method for monitoring the entire life cycle of cracks in concrete structures according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Call the crack displacement change, boundary offset direction sequence and width growth rate sequence corresponding to the active time period contained in the crack evolution active segment record, group the crack behaviors with continuous trends into the same group according to the monitoring area and time order, divide the groups according to the time series number, and generate crack behavior grouping index. S502: Based on the crack behavior grouping index, determine whether the offset direction of the crack remains consistent within the period, compare the width change trend within each group, filter the behavior set that simultaneously satisfies the consistency of direction and the continuous increase of width, and obtain the crack trend consistency marker group. S503: For the crack trend consistency marker group, determine that those that meet the trend conditions are classified as extended class, and the rest are classified as stable class. Summarize the category markers and time information of multiple groups, construct a structured record table with time period index and classification marker, and generate crack evolution classification label set.

10. A full life-cycle monitoring system for cracks in concrete structures, characterized in that, The system is used to implement the method for monitoring the entire life cycle of cracks in concrete structures as described in any one of claims 1-9, and the system comprises: The acoustic stress monitoring module is used to achieve S1: deploy acoustic points at the downstream dam pier nodes, measure the propagation time to establish a time series, calculate the rate of change of sound velocity, obtain the concrete stress change value, compare the stress curve with the fracture reference curve, and generate the crack initiation area signal. The crack contour recognition module is used to implement S2: acquire the inner wall image sequence of the crack initiation region signal, extract boundary feature points, calculate the Euclidean distance and angle change of the differential time, and if the direction is continuous and the distance increases, generate the crack contour expansion path. The crack synchronous propagation detection module is used to implement S3: Based on the crack profile propagation path, extract the crack length and width sequence of the main span bottom edge crack, calculate the growth rate, determine whether the two are rising synchronously and the difference is decreasing, and generate a crack synchronous propagation stage identifier. The crack evolution active cycle extraction module is used to implement S4: call the crack initiation region signal, the crack outline expansion path and the crack synchronous expansion stage identifier, perform intersection operation according to the timestamps of the three sets of results, filter the cycles that meet the active threshold, and generate crack evolution active segment records; The crack evolution classification module is used to implement S5: based on the records of the active crack evolution segments, analyze the changes in crack displacement, direction and width in time series, classify those with consistent direction and continuously increasing width into the extended class, and the rest into the stable class, and generate a crack evolution classification label set.

Images