A dam safety monitoring method and system

Abstract

The application discloses a dam safety monitoring method and system, relates to the technical field of hydraulic engineering, and comprises the following steps: three-dimensional arrangement of a controllable low-frequency excitation source and a distributed sensing array, active emission of a time sequence coded low-frequency elastic wave excitation signal, synchronous acquisition of response data and preprocessing; construction of a double-field coupling forward model based on an elastic wave wave equation and a seepage field equation, reconstruction of a three-dimensional wave velocity field and a permeability field by using a regularization iterative tomographic inversion algorithm; comparison of each element of a real-time field and a reference field, identification of a defect type and generation of a three-dimensional imaging atlas, realization of accurate positioning of the defect by relying on a ray tracing positioning algorithm, completion of quantitative grading early warning of the defect, early detection of hidden defects of the dam, global imaging, accurate positioning and grading early warning under the non-invasive condition of normal water storage of the reservoir and no structural damage.

Description

Technical Field

[0001] This invention relates to the field of water conservancy engineering technology, specifically to a method and system for dam safety monitoring. Background Technology

[0002] In water conservancy projects, dams serve as critical infrastructure, undertaking important functions such as flood control, irrigation, power generation, and water supply. However, during long-term operation, dams are inevitably affected by various factors such as water pressure, seepage, temperature changes, and geological settlement. This leads to the gradual appearance of hidden defects such as micro-leakage channels, internal loosening, crack expansion, and voids in the dam body, foundation, anti-seepage curtain, and structural contact interfaces. These defects are often highly concealed and develop rapidly in the early stages. If they are not detected and repaired in time, they may cause serious seepage control and structural instability problems, thereby threatening the safe operation of the dam. Therefore, early detection, full-area imaging, precise positioning, and graded early warning of hidden defects in dams have become important technical requirements for ensuring dam safety.

[0003] Existing dam safety monitoring technologies have shortcomings and cannot meet the high requirements of modern water conservancy projects for safety monitoring. First, traditional monitoring methods mainly rely on point sensors such as seepage pressure, strain, and displacement sensors, which have limited monitoring range and cannot form a full-area three-dimensional structural characterization of the dam body and foundation, resulting in blind spots in defect identification. Second, traditional monitoring relies heavily on passive response modes, depending on the structure's own response or changes in environmental loads, making it difficult to capture early, minute defects. As a result, defects are often discovered only after they have reached a certain scale, missing the best time for repair. Third, structural integrity monitoring and leakage monitoring are independent of each other, lacking a joint inversion mechanism of wave velocity field and permeability field, making it impossible to accurately distinguish defect types, resulting in low defect identification accuracy. Fourth, traditional detection methods often damage the dam structure, failing to meet the requirements of long-term online non-intrusive monitoring of dams in service, and the monitoring process has a significant impact on the normal operation of the dam. In addition, existing technologies also have problems such as response lag, imaging deficiencies, and inaccurate positioning, making it difficult to achieve dynamic management of hidden defects in the dam throughout its entire life cycle and failing to provide quantitative technical support for dam safety operation and maintenance. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method and system for dam safety monitoring. This system actively transmits time-coded low-frequency elastic wave excitation signals by deploying a controllable low-frequency excitation source and a distributed sensor array in three dimensions, simultaneously acquiring and preprocessing response data. Based on the elastic wave equation and seepage field equation, a dual-field coupled forward model is constructed, and a regularized iterative tomographic inversion algorithm is used to reconstruct the three-dimensional wave velocity field and permeability field. Through volume-by-volume comparison between the real-time field and the reference field, defect types are identified and a three-dimensional imaging atlas is generated. Relying on a ray tracing positioning algorithm, precise defect location is achieved, enabling quantitative and graded early warning of defects. This system can realize early detection, full-domain imaging, precise location, and graded early warning of hidden defects in dams under non-intrusive conditions of normal reservoir impoundment and no structural damage.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: On one hand, a method for monitoring the safety of a dam, comprising the following specific steps:

[0006] S1: In the early stage of dam completion or under structurally healthy conditions, the original response data of the dam body and foundation are obtained through active excitation and distributed acquisition, and the initial three-dimensional wave velocity reference field and three-dimensional permeability reference field are inverted and constructed.

[0007] S2: Multiple controllable low-frequency elastic wave excitation sources are arranged in preset spatial locations along the underwater area of ​​the dam's upstream side, the dam crest, the gallery, the curtain grouting area of ​​the dam foundation, and the dam shoulders on both banks.

[0008] S3: Deploy distributed sensor arrays inside the dam body, seepage prevention curtain, dam foundation, drainage gallery and rock foundation on both banks. Under the unified clock synchronization, each sensor unit collects the original vibration waveform data of the excitation wave propagating in the medium in real time, and extracts the vibration response, propagation delay, amplitude attenuation, phase shift and energy distribution data in the propagation path.

[0009] S4: Based on the elastic wave equation and the seepage field equation, a dual-field coupled forward model is constructed. The objective function is optimized by a regularized iterative inversion algorithm. After iterative convergence, the real-time three-dimensional wave velocity field and three-dimensional permeability field of the dam body and foundation are reconstructed.

[0010] S5: Compare the real-time three-dimensional wave velocity field, permeability field and reference field element by element, calculate the relative change rate of wave velocity and permeability, identify the defect type based on the preset threshold and abnormal combination features, and generate a three-dimensional structural defect imaging map through volume rendering technology.

[0011] S6: The ray tracing positioning algorithm combined with the three-dimensional mesh element coordinates is used to complete the three-dimensional precise positioning of defects. The severity level of defects is divided according to the abnormal amplitude, and the graded early warning is triggered and the positioning, type and level information are output.

[0012] S7: Locate, quantitatively assess, and classify defects for early warning, and execute the process cyclically to achieve long-term online monitoring.

[0013] Furthermore, in S1, during the initial stage of dam completion or when the structure is in a healthy state, a three-dimensional distributed sensing array is used to collect the original data of the time-domain response, propagation delay, amplitude attenuation, and phase shift of the excitation wave propagating in the dam body, dam foundation, and anti-seepage curtain medium under high-precision clock synchronization. After preprocessing the collected data by noise reduction filtering, effective signal acquisition, and amplitude and phase correction, the elastic wave equation is used as the forward calculation model of the wave field, and the continuity and motion equation of the seepage field are used as parameter constraints. The three-dimensional space of the dam body and foundation is finely meshed, and the iterative inversion algorithm is used to calculate the medium wave velocity and permeability distribution parameters element by element. Through multiple rounds of excitation acquisition and iterative optimization, measurement errors are eliminated, and finally, an initial three-dimensional wave velocity reference field and a three-dimensional permeability reference field covering the entire dam main structure and foundation are constructed.

[0014] Furthermore, in S4, the elastic wave equation serves as the basis for the forward calculation of the wave field. A dual-field coupled forward model is constructed by coupling the continuity equation and the motion equation of the seepage field. Real-time response data collected by the distributed sensor array is used as the inversion constraint. A regularized iterative tomographic inversion algorithm is employed to reconstruct the parameter field. The inversion objective function is used as the core of iterative optimization. The parameter field of the baseline model under the dam's healthy state is used as the initial iteration value. The wave velocity and permeability model parameters are updated round by round using the gradient descent method. The objective function value is calculated after each iteration until the rate of change of the objective function is ≤10. -6 The convergence conditions are met, and the real-time reconstruction of the three-dimensional wave velocity field and the three-dimensional permeability field is completed.

[0015] Furthermore, in step S4, a regularized iterative tomographic inversion algorithm is used to reconstruct the parameter field in order to invert the objective function. As the core of iterative optimization, the baseline model parameter field under the dam's healthy state is used. Using the initial iteration values, the wave velocity and permeability model parameters are updated round by round using the gradient descent method. The objective function value is calculated after each iteration until the rate of change of the objective function is ≤10. -6 The convergence condition is satisfied, where, It is the inversion objective function, used to measure the deviation between the model's calculated values ​​and the measured values, as well as the degree of model constraint. It is the model parameter field to be inverted, which includes wave velocity and permeability parameters in the three-dimensional space of the dam body and foundation; It is the observation data weight matrix, used to balance the observation weights of response data from different sensor channels and different locations; It is the excitation response data obtained from the actual measurement of the distributed sensing array, which includes information on wave propagation delay, amplitude, phase and energy attenuation; It is a forward model operator for elastic wave-seepage field coupling, which characterizes the mapping relationship between model parameters and excitation response data; It is a regularization coefficient used to suppress inappropriate issues during the inversion process. It is the model parameter weight matrix, used to constrain the spatial smoothness and physical rationality of the wave velocity field and the permeability field; It is the baseline model parameter field under the healthy state of the dam, namely the initial three-dimensional wave velocity baseline field and permeability baseline field.

[0016] Furthermore, in step S5, the real-time three-dimensional wave velocity field and three-dimensional permeability field are compared one by one with the initial three-dimensional wave velocity reference field and three-dimensional permeability reference field under the dam's healthy state. The real-time wave velocity, real-time permeability, and corresponding reference values ​​of each three-dimensional grid cell are extracted. The abnormal amplitude of each cell is calculated using the wave velocity relative change rate formula and the permeability relative change rate formula. A unified and adaptively fine-tunable anomaly detection threshold is set, wherein the basic anomaly threshold is set as: wave velocity relative change rate Relative change rate of penetration 3D mesh elements that meet this condition are classified as structurally abnormal elements, while elements that do not meet this threshold are classified as healthy elements. Based on this, according to and By combining the abnormal amplitude combination characteristics with the differences in the physical properties of dam defects, the type of structural defect can be accurately determined: if and It was determined to be a tiny leakage channel; if Between 10% and 20% If the area is between 10% and 30%, it is considered an internal loose zone; if the local volume element... Surrounding elements No obvious abnormalities and If it is determined to be a crack or void defect, and Then it is determined to be a healthy area.

[0017] Furthermore, in step S5, the abnormal amplitude of each volume element is calculated using the formulas for the relative rate of change of wave velocity and the relative rate of change of permeability, wherein the formula for the relative rate of change of wave velocity is: The formula for the relative rate of change of penetration is: In the formula, The relative rate of change of wave velocity is used to determine the degree of abnormality in the compactness of the dam medium and the structural integrity. The medium wave velocity value obtained through real-time inversion; This is the reference value for the medium wave velocity under healthy conditions; It is the relative rate of change of permeability, used to determine the seepage characteristics of the dam body and the degree of abnormality of the seepage channels; It is the medium permeability value obtained by real-time inversion; It is the benchmark value of media permeability under healthy conditions.

[0018] Furthermore, in S6, a ray tracing positioning algorithm combined with the spatial coordinate information of three-dimensional mesh elements is used to accurately locate the marked structural anomaly area in three-dimensional space. First, a global geodetic coordinate system for the dam is established, and the spatial coordinates of the three-dimensional mesh elements are precisely calibrated with the actual structural position of the dam. Taking each controllable low-frequency excitation source as the ray emission starting point and the corresponding distributed sensing unit receiving the excitation response signal as the ray receiving endpoint, the propagation path of each ray in the dam body and foundation medium is calculated by combining the three-dimensional wave velocity field parameters obtained by real-time inversion, and the refraction and reflection errors of elastic waves in different media are corrected. Ray coverage analysis is performed on the marked anomaly elements, and all ray beams penetrating the anomaly element are counted. By fitting the intersection coordinates of multiple ray beams and spatial interpolation calculation, the accurate three-dimensional spatial coordinates of the anomaly element are determined, eliminating the deviation of single ray positioning. For continuously distributed anomaly elements, the spatial extension trajectory, length, and cross-sectional dimensions of the defect are fitted by the continuity analysis of the ray tracing path. For discrete and abruptly changing anomaly elements, the center point coordinates and influence range of the defect are determined by focusing the intersection of ray beams, and finally, the three-dimensional spatial positioning of the defect is achieved.

[0019] Furthermore, in S6, based on the relative rate of change of wave speed... Relative change rate of permeability Based on the abnormal amplitude, combined with the defect's spatial range and extension length, the severity of the defect is divided into four levels:

[0020] Level I: Between 5% and 10% Between 10% and 20%, the defect scale is small and there is no obvious development trend;

[0021] Level II: Between 10% and 20% Between 20% and 50%, the defects have reached a certain scale and require regular monitoring.

[0022] Level III: Between 20% and 30%, Between 50% and 80%, defects show a growing trend and are prone to causing safety hazards;

[0023] Level IV: , The defect has affected the structural stability and needs to be addressed immediately.

[0024] Furthermore, in S6, the determination results of defect type and abnormality degree are associated with each volume element in the reconstructed three-dimensional wave velocity field and three-dimensional permeability field. Attribute annotation is performed on each three-dimensional mesh volume element. Combined with volume rendering technology, different color gradients and textures are used to distinguish healthy areas, various defects and different abnormalities, ultimately forming a high-resolution three-dimensional real-time parameter field.

[0025] On the other hand, a dam safety monitoring system includes:

[0026] Controllable low-frequency excitation module: performs time-sequence encoding and time-division transmission of controllable low-frequency elastic wave excitation signal, adaptively adjusts relevant signal parameters, and cooperates with the calibration signal of the transmission system to support the establishment of the reference field;

[0027] Distributed sensing acquisition module: performs multi-channel synchronous acquisition and preprocessing of excitation wave response data, providing data support for dual-field coupling inversion and ray tracing positioning;

[0028] Clock synchronization module: It adopts a dual-mode clock synchronization strategy to provide a high-precision unified time reference and ensure the time consistency of excitation transmission, data acquisition and positioning calculation.

[0029] Data transmission module: It adopts a composite transmission method to adapt to different monitoring scenarios, realizes data transmission, and has data caching and encryption functions;

[0030] Tomographic imaging processing unit: integrates data processing, dual-field coupling inversion, three-dimensional imaging, defect identification, ray tracing positioning and data storage functions, and completes the core processing of monitoring data and defect judgment;

[0031] Upper computer monitoring and early warning unit: realizes the visualization of monitoring data, three-dimensional imaging map and defect information, supports historical data query, trend analysis and automatic report generation, and completes defect classification early warning and remote system control.

[0032] Compared with existing technologies, this dam safety monitoring method and system has the following advantages:

[0033] This invention employs an active low-frequency excitation mode to stimulate the weak response of early-stage minor defects in dam structures. Combined with dual-field coupling inversion technology, it can capture early-stage hidden defects that traditional passive monitoring cannot detect, enabling early detection and early warning of potential hazards and significantly improving the proactiveness of dam safety management. Through joint inversion of three-dimensional wave velocity and three-dimensional permeability fields, combined with the abnormal amplitude combination characteristics of the relative change rate of wave velocity and permeability, it can accurately distinguish different types of defects such as seepage channels, internal looseness, cracks, and voids, avoiding misjudgments caused by single-parameter inversion. The defect identification accuracy is high. By integrating ray tracing positioning algorithms and combining three-dimensional grid element coordinate calibration and wave velocity parameter correction, it can achieve precise three-dimensional spatial positioning of defects, accurately locking the location, extension trajectory, and impact range of defects, providing precise targeting for dam maintenance and reinforcement, and reducing operation and maintenance costs. At the same time, the three-dimensional grid is used to deploy excitation sources and sensor arrays to achieve full-area monitoring of the dam body, dam foundation, and anti-seepage curtain without blind spots. The monitoring process is executed cyclically, and a historical database is established to trace the development trend of defects, realizing dynamic management of hidden defects throughout their entire life cycle.

[0034] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0036] Figure 1 A flowchart of a dam safety monitoring method;

[0037] Figure 2 A flowchart of step S4 in a dam safety monitoring method;

[0038] Figure 3 This is a structural block diagram of a dam safety monitoring system;

[0039] Figure 4 This is a structural block diagram of a tomographic imaging processing unit in a dam safety monitoring system. Detailed Implementation

[0040] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0041] This invention provides a method and system for dam safety monitoring. It actively transmits time-coded low-frequency elastic wave excitation signals by deploying a three-dimensional controllable low-frequency excitation source and a distributed sensor array, simultaneously acquiring and preprocessing response data. A dual-field coupled forward model is constructed based on the elastic wave equation and the seepage field equation. A regularized iterative tomographic inversion algorithm is used to reconstruct the three-dimensional wave velocity field and permeability field. By comparing the real-time field with the reference field element-by-element, defect types are identified and a three-dimensional imaging atlas is generated. A ray-tracing localization algorithm is used to achieve precise defect location and quantitative, graded early warning of defects. This system can achieve early detection, full-domain imaging, precise location, and graded early warning of hidden defects in dams under non-intrusive conditions of normal reservoir impoundment and no structural damage.

[0042] like Figure 1 As shown, S1: In the early stage of dam completion or under structurally healthy conditions, the original response data of the dam body and foundation are obtained through active excitation and distributed acquisition, and the initial three-dimensional wave velocity reference field and three-dimensional permeability reference field are constructed by inversion.

[0043] In the early stages of dam completion or when the structure is confirmed to be in a healthy and stable state after testing, a controllable low-frequency elastic wave excitation signal of 0.1-50Hz is emitted to the entire dam body, dam foundation, and anti-seepage curtain. A three-dimensional distributed sensor array is used to collect the original data of the time domain response, propagation delay, amplitude attenuation, and phase shift of the excitation wave in different media such as concrete, rock foundation, and grouting body under nanosecond-level high-precision clock synchronization. The collection range covers the main structure of the dam, the entire dam foundation, the anti-seepage curtain, and the rock foundation of the dam abutments on both banks. During the collection process, the coverage of the excitation wave propagation path in each monitoring area is ensured to be ≥95%, and the original data sampling rate is set to 1kHz-10kHz to ensure that the time domain resolution meets the parameter calculation requirements.

[0044] The collected raw data undergoes multiple preprocessing steps to remove invalid signals and correct system errors, ensuring data accuracy.

[0045] Noise reduction filtering: Wavelet noise reduction algorithm + finite impulse response (FIR) filtering is used to remove invalid signals such as environmental vibration, equipment noise, and electromagnetic interference, while retaining the effective response signal of the excitation wave propagation. The signal-to-noise ratio of the filtered signal is ≥30dB.

[0046] Effective signal acquisition: The threshold method and cross-correlation method are used to extract the characteristic signals of the excitation wave arriving at each sensing unit, accurately identify the wavefront arrival time, and ensure that the propagation delay calculation error is ≤1μs;

[0047] Amplitude and phase correction: Based on the factory calibration curve of the sensing unit and the field calibration results, the signal amplitude is linearly corrected, and the phase deviation of different propagation paths is corrected by the linear phase compensation method.

[0048] Unstructured tetrahedral meshes were used to divide the dam body and foundation into three dimensions. The meshing accuracy was determined based on the dam's structural characteristics and key monitoring areas.

[0049] Key monitoring areas (underwater area on the water-facing side, dam foundation curtain grouting area, dam abutment junction on both banks, and corridor perimeter): volume element size is set to 0.5-2m to ensure local monitoring accuracy;

[0050] For general monitoring areas (middle of the dam body, non-seepage-proof area of ​​the dam foundation): the volume element size is set to 2-5m to improve the efficiency of subsequent calculations while ensuring the monitoring effect;

[0051] After being partitioned, a three-dimensional mesh set of voxels covering the entire dam area is formed. Each voxel is assigned a unique spatial coordinate code to achieve a one-to-one correspondence with the actual structural location of the dam.

[0052] Using the elastic wave equation of viscoelastic medium as the forward wave field calculation model, and Darcy's law combined with the continuity of the seepage field and the equation of motion as parameter constraints, a conjugate gradient iterative inversion algorithm is used to calculate the medium wave velocity and permeability distribution parameters on a volumetric basis. Through 3-5 rounds of active excitation and data acquisition, combined with multiple rounds of iterative optimization, random and systematic errors in the measurement process are eliminated. After each round of iteration, the consistency of the parameter results is verified, and data is re-acquired when the deviation exceeds 5%. Finally, an initial three-dimensional wave velocity reference field and a three-dimensional permeability reference field covering the entire dam structure and foundation are constructed. The reference field stores the healthy state wave velocity reference value corresponding to each three-dimensional grid volume element. Compared with the permeability benchmark value And establish a reference field database to achieve permanent data storage.

[0053] S2: Multiple controllable low-frequency elastic wave excitation sources are arranged in preset spatial locations along the underwater area of ​​the dam's upstream side, the dam crest, the gallery, the curtain grouting area of ​​the dam foundation, and the dam shoulders on both banks.

[0054] Multiple controllable low-frequency elastic wave excitation sources are arranged along the key structural areas of the dam according to the principles of full coverage, key reinforcement, and spatial uniformity. The location and density of the excitation sources are adapted to the dam structure size to ensure that the excitation waves can cover the entire dam monitoring area.

[0055] Arrangement area and spacing:

[0056] The underwater area of ​​the dam's upstream face is divided into layers according to a water depth of 10-20m, with each layer arranged at a horizontal interval of 30-50m. The underwater excitation source has waterproof and water pressure resistant characteristics and is suitable for underwater working environments with a water depth of 0-50m.

[0057] Dam crest: Along the dam axis, the excitation sources are arranged at equal intervals of 50-100m, and are installed on the monitoring piers on the dam crest, which are rigidly connected to the dam body to ensure effective propagation of the excitation wave;

[0058] Corridors: Arranged at intervals of 20-30m along the corridor extension direction, with excitation sources added at corridor turns and cross-section changes;

[0059] The grouting zone of the dam foundation is arranged according to the pattern of grouting hole row spacing of 3-5m and hole spacing of 5-8m, forming a spatial match with the grouting hole positions;

[0060] The abutments on both sides of the dam are arranged according to the degree of weathering of the bedrock. Slightly weathered bedrock is arranged at 50m intervals, and weakly / strongly weathered bedrock is arranged at 30m intervals, covering the key parts where the abutments connect to the dam body.

[0061] Excitation source function configuration:

[0062] All excitation sources have independent signal transmission control capabilities, enabling time-division, zone-division, and frequency-division transmission of excitation signals to avoid mutual interference between multiple excitation source signals; the excitation source can transmit low-frequency elastic waves of 0.1-50Hz, with the output amplitude adaptively adjusted within the range of 0-10V to meet the propagation requirements of excitation waves in different media; the excitation source establishes a communication connection with the tomographic imaging processing unit and receives remote control commands to achieve real-time adjustment of transmission parameters.

[0063] S3: Deploy distributed sensor arrays inside the dam body, seepage prevention curtain, dam foundation, drainage gallery and rock foundation on both banks. Under the unified clock synchronization, each sensor unit collects the original vibration waveform data of the excitation wave propagating in the medium in real time, and extracts the vibration response, propagation delay, amplitude attenuation, phase shift and energy distribution data in the propagation path.

[0064] Distributed sensor arrays are deployed inside the dam body, the seepage prevention curtain, the dam foundation, the drainage gallery, and the bedrock on both banks. The sensing units of the sensor array are spatially matched with the controllable low-frequency elastic wave excitation source to ensure full-dimensional coverage of the excitation wave propagation path.

[0065] The combination of MEMS vibration sensor and fiber optic acoustic sensor can accurately collect the vibration response and acoustic propagation characteristics of the excitation wave, and is suitable for complex working environments such as inside the dam body, underwater, and rock foundation. The deployment density is adapted to the three-dimensional grid division accuracy. The sensor units in key monitoring areas are deployed at a spacing of 1-3m, and the general monitoring areas are deployed at a spacing of 5-8m.

[0066] All sensing units are connected to a unified clock synchronization module to achieve nanosecond-level high-precision clock synchronization with a clock deviation of ≤1ns, ensuring the consistency of data acquisition time for each sensing unit and providing a time reference for the accurate calculation of parameters such as excitation wave propagation delay and phase shift.

[0067] After the controllable low-frequency elastic wave excitation source transmits the excitation signal according to the preset command, each sensing unit of the distributed sensing array collects the vibration response, propagation delay, amplitude attenuation, phase shift and energy distribution data of the excitation wave in the propagation path in real time and synchronously. The collected raw data is initially filtered on site and then uploaded to the data transmission module in real time. During the acquisition process, the data quality is monitored in real time. If data loss or signal interruption occurs, a re-acquisition command is immediately triggered to ensure data integrity ≥99%.

[0068] like Figure 2 As shown, S4: A dual-field coupled forward model is constructed based on the elastic wave equation and the seepage field equation. The objective function is optimized by a regularized iterative inversion algorithm. After iterative convergence, the real-time three-dimensional wave velocity field and three-dimensional permeability field of the dam body and foundation are reconstructed.

[0069] Based on the elastic wave equation of viscoelastic medium as the basis for forward wave field calculation, a dual-field coupled forward model is constructed by coupling the continuity equation of seepage field and the motion equation of Darcy's law. This model can accurately characterize the intrinsic mapping relationship between the wave velocity parameters, permeability parameters and excitation wave response data of dam body and foundation medium. At the same time, it considers the influence of medium heterogeneity and anisotropy on elastic wave propagation and seepage characteristics, thus improving the model's fit.

[0070] The real-time vibration response, propagation delay, amplitude attenuation and other excitation response data collected by the distributed sensing array are used as hard constraints for inversion. The parameter field is reconstructed by a regularized iterative tomographic inversion algorithm. This algorithm can effectively suppress ill-posed problems in the inversion process and ensure the physical rationality and spatial smoothness of the inversion results.

[0071] Inverting the objective function The core of the iterative optimization is the parameter field of the model to be inverted. The wave velocity and permeability parameters, containing all three-dimensional mesh elements across the entire dam area, are stored in a two-dimensional matrix format; the observation data weight matrix... Assigning values ​​based on the signal-to-noise ratio (SNR) of the sensing units: Sensing units with a high SNR (≥30dB) are assigned a weight of 1-1.5, those with a low SNR (20-30dB) are assigned a weight of 0.5-1, and those with no valid signal are assigned a weight of 0, thus achieving weight balance for data from different sensing channels; measured excitation response data. Multi-dimensional data acquired and preprocessed by a distributed sensing array, including wave propagation delay, amplitude, phase, and energy attenuation information; coupled with a forward model operator. The numerical solution operator for the two-field coupled forward model is obtained by the finite element method; regularization coefficients. The value ranges from 0.001 to 0.1, and it is adaptively adjusted according to the noise level of the monitored data; the higher the noise level, the better. The larger the value, the more effectively the oscillation problem is suppressed during the inversion process; model parameter weight matrix A spatial smoothing matrix is ​​used to constrain the parameter variations of adjacent volume elements, avoiding physically meaningless abrupt changes in the inversion results and ensuring the spatial continuity of the wave velocity field and permeability field; the baseline model parameter field The initial three-dimensional wave velocity and permeability reference fields under the healthy state of the dam are used as the initial values ​​for iteration.

[0072] Based on the baseline model parameter field Using the initial iteration values, the wave velocity and penetration rate model parameters are updated round by round using the adaptive learning rate gradient descent method. The learning rate is set to 0.0001-0.001 and adaptively decreases with the number of iterations; the objective function is calculated after each iteration. The value is set, and dual convergence criteria are set:

[0073] Core condition: The rate of change of the objective function ≤ 10 -6 ;

[0074] Auxiliary condition: Maximum number of iterations ≤ 1000;

[0075] The iteration stops when any condition is met, and the real-time three-dimensional wave velocity field and three-dimensional permeability field of the dam body and foundation are reconstructed. The real-time field after reconstruction stores the real-time wave velocity value and real-time permeability value corresponding to each three-dimensional grid element. The relative error of the reconstruction result is ≤3%.

[0076] S5: Compare the real-time three-dimensional wave velocity field, permeability field and reference field element by element, calculate the relative change rate of wave velocity and permeability, identify the defect type based on the preset threshold and abnormal combination features, and generate a three-dimensional structural defect imaging map through volume rendering technology.

[0077] The reconstructed real-time three-dimensional wave velocity field and permeability field are compared with the constructed reference field on a volume-by-volume basis. Structural anomalies are identified by calculating the relative rate of change of parameters. The defect type is determined by combining the characteristics of anomaly combinations. High-resolution three-dimensional structural defect imaging atlases are generated based on volume rendering technology.

[0078] Parameters are extracted from the same three-dimensional mesh volume elements in the real-time and reference fields to obtain real-time wave velocity values, real-time permeability values, and corresponding reference values. The relative rate of change of wave velocity for each volume element is then calculated using the following formulas. Relative change rate of permeability The calculation result should be rounded to two decimal places.

[0079] Relative rate of change of wave speed: ;

[0080] Relative rate of change in penetration: ;

[0081] in, Used to determine the degree of abnormality in the compactness of the dam body medium and the structural integrity. Used to determine the seepage characteristics of the dam body and the degree of abnormality of the seepage channels.

[0082] Set a uniform and adaptively fine-tunable basic threshold for anomaly detection: relative rate of change of wave velocity. Relative change rate of penetration 3D mesh elements that meet this basic threshold are classified as structurally abnormal elements, while elements that do not meet this threshold are classified as healthy elements.

[0083] The threshold can be adaptively fine-tuned according to the dam type, operating years, and hydrogeological conditions. For example, the wave velocity threshold for concrete dams can be fine-tuned to 4%, and the permeability threshold can be fine-tuned to 8%; the basic threshold for earth-rock dams remains unchanged; for dams that have been in operation for more than 10 years, the threshold can be increased by 1%-2% to adapt to the monitoring needs of dam structural aging.

[0084] in accordance with and Based on the abnormal amplitude combination characteristics and the differences in the physical properties of dam defects, the type of structural defect is accurately determined. The determination rules strictly match the physical change law of the dam medium. The specific determination criteria are as follows:

[0085] Micro-leakage channels: and This type of defect manifests as a significant decrease in medium wave velocity and a sharp increase in permeability, resulting in interconnected seepage channels in the dam body medium.

[0086] Internal loose area: Between 10% and 20% Between 10% and 30%, this type of defect manifests as a decrease in the density of the medium, an increase in porosity, and a slight change in seepage characteristics, corresponding to loose concrete in the dam body and the early stage of rock foundation fissure development.

[0087] Cracks or voids: local volume elements Surrounding elements No obvious abnormalities and This type of defect manifests as local structural integrity failure with no significant change in seepage characteristics, corresponding to surface / internal cracks in the dam body and voids at the junction of the dam body and the bedrock.

[0088] Health area: and The wave velocity and permeability of the medium did not change significantly, and the structure was in a healthy state.

[0089] The results of the defect type and abnormal amplitude determination are associated with each element in the reconstructed real-time three-dimensional wave velocity field and three-dimensional permeability field. Attribute annotations are performed on each three-dimensional mesh element (annotation content: healthy / abnormal, defect type, relative change rate).

[0090] A ray casting method for volume rendering is employed, and a 3D imaging model is constructed based on the VTK visualization engine. Different color gradients and texture density are used to distinguish healthy areas, various defects, and different levels of anomalies. Specific visual settings are as follows:

[0091] Color gradient: healthy areas are blue, internal loose areas are yellow, cracks / void defects are orange, and tiny leakage channels are red; the greater the abnormality, the higher the color saturation.

[0092] Texture density: Anomalies ≤10% indicate sparse texture, 10%-30% indicate medium texture, and ≥30% indicate dense texture;

[0093] The final result is a high-resolution three-dimensional structural defect imaging atlas with a resolution of ≥0.5m, which can be magnified, reduced, rotated, and sectioned to intuitively present the distribution, anomaly amplitude, and spatial location of structural defects throughout the dam area.

[0094] S6: The ray tracing positioning algorithm combined with the three-dimensional mesh element coordinates is used to complete the three-dimensional precise positioning of defects. The severity level of defects is divided according to the abnormal amplitude, and the graded early warning is triggered and the positioning, type and level information are output.

[0095] Using the geodetic coordinate system from the dam engineering survey as a reference, the defects were precisely located in three dimensions with a positioning error ≤ 0.5m. Specific operational steps are as follows:

[0096] Establish a global geodetic coordinate system for the dam, and accurately calibrate the spatial coordinates of the subdivided three-dimensional mesh elements with the actual structural position of the dam. The calibration error is ≤0.1m, achieving a one-to-one correspondence between the element coordinates and the actual position of the dam.

[0097] Using each controllable low-frequency elastic wave excitation source as the starting point of ray emission and the corresponding distributed sensing unit receiving the excitation response signal as the ending point of ray reception, combined with the real-time three-dimensional wave velocity field parameters obtained by inversion, the propagation path of each ray in different media of the dam body and dam foundation is calculated by the fast travel method of the process function equation. At the same time, the refraction and reflection errors of elastic waves at different media interfaces are corrected according to Snell's law to ensure the accuracy of the propagation path calculation.

[0098] Ray coverage analysis is performed on the marked structural anomalous elements. An effective coverage threshold is set: an anomalous element is considered effective if at least 3 rays penetrate it. All ray rays that penetrate effective anomalous elements are counted. The coordinates of the intersection points of multiple rays are fitted by Kriging space interpolation to determine the accurate three-dimensional spatial coordinates of the anomalous elements and eliminate the deviation of single ray positioning.

[0099] For continuously distributed anomalous elements (such as extended cracks and leakage channels), the spatial extension trajectory, length, and cross-sectional dimensions of the defects are obtained by fitting the spline interpolation method through continuity analysis of the ray tracing path. For discrete anomalous elements (such as local voids and point cracks), the intersection of the focused ray beams is fitted by the least squares method to determine the coordinates of the center point of the defect and its influence range. Finally, the three-dimensional spatial positioning of the defects is accurately achieved, and a defect positioning coordinate table is generated.

[0100] Based on the relative rate of change of wave speed Relative change rate of permeability Based on the abnormal amplitude, combined with the spatial range, extension length, and distribution location of the defects (with increased weighting for defects in key areas), the severity of defects is divided into four levels. The level classification fully considers the degree of impact of defects on the stability of the dam structure. The specific classification criteria are as follows:

[0101] Level I: Between 5% and 10% Between 10% and 20%, the defect scale is small and there is no obvious development trend;

[0102] Level II: Between 10% and 20% Between 20% and 50%, the defects have reached a certain scale and require regular monitoring.

[0103] Level III: Between 20% and 30%, Between 50% and 80%, defects show a growing trend and are prone to causing safety hazards;

[0104] Level IV: , The defect has affected the structural stability and needs to be addressed immediately.

[0105] A graded early warning mechanism is triggered based on the severity level of the defect. Early warning information, along with data such as defect location, type, and severity, is output in real time. Specific early warning rules and information output requirements are as follows:

[0106] Warning levels: Level I triggers a yellow warning, Level II triggers an orange warning, Level III triggers a red warning, and Level IV triggers a red emergency warning;

[0107] Warning methods: on-site audible and visual alarms, PC pop-up alarms, mobile terminal (SMS / APP) push notifications, and email push notifications, with multiple terminals triggering simultaneously to ensure timely delivery of warning information;

[0108] Information output: The output includes the three-dimensional spatial coordinates of the defect, defect type, severity level, impact range, abnormal amplitude, and development trend prediction. It also includes a screenshot of the three-dimensional imaging atlas of the defect area, providing comprehensive and accurate decision-making basis for the safe operation and maintenance of the dam.

[0109] S7: Locate, quantitatively assess, and classify defects for early warning, and execute the process cyclically to achieve long-term online monitoring;

[0110] The monitoring cycle is dynamically adjusted based on the dam's operational status, hydrological conditions, and meteorological conditions to ensure both real-time monitoring and cost-effectiveness. Specific cycle setting standards are as follows:

[0111] Under normal operating conditions (outside of the flood season, when water levels are stable and the weather is calm): the monitoring cycle is 24 hours / time;

[0112] Special operating conditions (flood season, high water level operation, post-earthquake, before / after rainstorm): The monitoring cycle is adjusted to 1 hour / time to achieve high-frequency real-time monitoring;

[0113] Extreme operating conditions (abnormal dam deformation, discovery of Class III / IV defects, torrential rain): The monitoring cycle is adjusted to 15 minutes / time to achieve ultra-high frequency continuous monitoring.

[0114] When the dam experiences the following conditions, the initial three-dimensional wave velocity reference field and the three-dimensional permeability reference field are dynamically corrected to ensure the effectiveness of the reference fields and the accuracy of the monitoring results:

[0115] After the dam has undergone repair, reinforcement, and seepage prevention engineering measures;

[0116] After the dam has been in operation for more than 10 years or after a moderate or greater earthquake (magnitude ≥ 4.0);

[0117] When the overall deviation between the reference field and the real-time field continues to exceed 10%;

[0118] Correction method: Re-implement active incentive, data collection and iterative inversion work, combine historical baseline data for weighted fusion, construct a new baseline field to replace the original baseline field as the reference for subsequent monitoring;

[0119] All monitoring data (raw data, real-time field data, defect identification and location results, and early warning information) are permanently stored in a distributed database with a storage period of ≥10 years. At the same time, time series trend analysis is performed on historical monitoring data to predict the development trend of defects, providing data support for the dam's full life cycle safety assessment and operation and maintenance planning.

[0120] like Figure 3 As shown, the present invention also provides a dam safety monitoring system, which is adapted to the above-mentioned dam safety monitoring method and can realize the full-process functions such as excitation signal transmission, response data acquisition, data processing, defect identification and location, and graded early warning. The system includes a controllable low-frequency excitation module, a distributed sensor acquisition module, a clock synchronization module, a data transmission module, a tomographic imaging processing unit, and a host computer monitoring and early warning unit.

[0121] The controllable low-frequency excitation module provides elastic wave excitation signals for dam safety monitoring and is the core module for realizing active monitoring. It mainly completes the generation, encoding, transmission, and parameter adjustment of controllable low-frequency elastic wave excitation signals. Its specific functions are as follows:

[0122] The built-in excitation signal generation unit can generate a controllable low-frequency elastic wave excitation signal that meets the monitoring requirements. The frequency, amplitude, waveform and other parameters of the excitation signal can be adjusted according to the monitoring scenario.

[0123] The excitation signal is processed by timing encoding, and the time-division multiplexing method is used to realize the independent transmission control of each excitation source, so as to avoid mutual interference between the signals of multiple excitation sources;

[0124] It has the ability to adaptively adjust signal parameters, and can adaptively adjust parameters such as frequency, amplitude, and transmission duration of the excitation signal according to the structural characteristics of the dam and the medium characteristics of the monitoring area, so as to ensure the propagation effect of the excitation signal and the effectiveness of data acquisition.

[0125] It can be used in conjunction with the calibration signal of the transmission system to transmit standard calibration signals during the construction of the reference field, providing precise excitation signal support for the construction of the reference field;

[0126] It receives control commands from the tomographic imaging processing unit to transmit excitation signals in a time-division and zone-division manner, thus coordinating with the distributed sensing acquisition module.

[0127] The distributed sensing acquisition module mainly completes multi-channel synchronous acquisition of excitation wave response data, signal preprocessing, and preliminary data extraction, providing accurate and effective data support for subsequent dual-field coupling inversion and ray tracing positioning. Its specific functions are as follows:

[0128] The distributed sensing array consists of several sensing units. The sensing units can accurately collect data such as vibration response, propagation delay, amplitude attenuation, phase shift and energy distribution during the propagation of the excitation wave. It has the ability to acquire data simultaneously through multiple channels, and the number of acquisition channels can be expanded according to monitoring needs.

[0129] The built-in signal conditioning and preprocessing unit performs preprocessing such as filtering, amplification, amplitude and phase correction on the acquired raw signal to remove invalid noise signals and improve the accuracy and effectiveness of the data.

[0130] It can extract the feature parameters of the response data in real time and synchronously upload the preprocessed raw data and feature parameters to the data transmission module.

[0131] It receives the clock synchronization signal from the clock synchronization module to achieve high-precision time synchronization of all sensing units and ensure the time consistency of data acquisition.

[0132] It features low power consumption and anti-interference operation characteristics, and can adapt to complex working environments such as inside the dam body, underwater, and rock foundation.

[0133] The clock synchronization module employs a dual-mode clock synchronization strategy, providing a high-precision unified time reference to ensure time consistency between excitation transmission, data acquisition, and positioning calculation. Its specific functions include:

[0134] It integrates two clock synchronization modes, which can be switched according to the monitoring environment and communication conditions of the dam. In areas with good satellite signal coverage, the satellite synchronization mode is used to achieve high-precision clock synchronization; in areas where satellite signals are blocked, such as corridors and inside the dam body, the local crystal oscillator synchronization mode is used, and the synchronization accuracy of the local crystal oscillator is ensured by combining the calibration signal of satellite synchronization.

[0135] It can output a unified clock synchronization signal to the controllable low-frequency excitation module, the distributed sensing acquisition module, and the tomographic imaging processing unit to achieve time synchronization of each module / unit. The synchronization accuracy meets the calculation requirements of parameters such as excitation wave propagation delay and phase shift.

[0136] It has clock skew self-correction capability, which can monitor the clock skew of each module / unit in real time and perform automatic correction to ensure the consistency and stability of system time.

[0137] The data transmission module adopts a composite transmission method to adapt to different monitoring scenarios, realizes data transmission, and has data caching and encryption functions. The specific functions are as follows:

[0138] Integrating multiple data transmission methods, a wired + wireless composite transmission system is formed. In areas where wired communication can be laid, such as inside the dam body and in corridors, wired transmission is used to ensure the stability and high speed of data transmission. In areas where wired communication is difficult to lay, such as underwater on the water-facing side and on both banks of the dam, wireless transmission is used to realize wireless data transmission and reception.

[0139] It has the capability to transmit large amounts of data at high speed, and can transmit massive amounts of response data collected by the distributed sensing acquisition module in real time, while also transmitting control commands and status information between modules / units.

[0140] The built-in data caching unit can cache the collected data locally in case of network interruption, transmission failure or other abnormal situations, and retransmit it after communication is restored to avoid data loss.

[0141] It has data encryption capabilities and uses encryption algorithms to encrypt the transmitted data, ensuring the security and confidentiality of the monitoring data and preventing the data from being tampered with or stolen.

[0142] It can classify and forward transmitted data, forward the collected response data to the tomographic imaging processing unit, and forward the early warning information and monitoring data to the host computer monitoring and early warning unit.

[0143] like Figure 4 As shown, the tomographic imaging processing unit integrates full-process processing functions such as data processing, dual-field coupling inversion, three-dimensional imaging, defect identification, ray tracing localization, and data storage. Specific functions include:

[0144] The built-in data processing subunit performs secondary processing on the preprocessed data uploaded by the distributed sensor acquisition module, including data normalization, outlier removal, and data interpolation, to further improve the quality of the data.

[0145] It has a built-in dual-field coupling inversion sub-unit, which integrates the elastic wave equation, seepage field equation and regularized iterative tomographic inversion algorithm. It can complete the reconstruction of the real-time three-dimensional wave velocity field and three-dimensional permeability field, and store the benchmark model parameter field under the dam health state.

[0146] With a built-in 3D imaging subunit and integrated volume rendering technology, it can generate high-resolution 3D structural defect imaging maps based on defect identification results, realizing 3D visualization of defect distribution;

[0147] It has a built-in defect identification and location subunit, which integrates a calculation model for the relative change rate of wave velocity and permeability, defect type determination rules and ray tracing location algorithm, and can complete the identification, three-dimensional precise location and severity level classification of structural defects;

[0148] The built-in data storage subunit can store raw collected data, processed data, reference field data, real-time field data, defect identification and location results for a long time, providing data support for historical data query and trend analysis;

[0149] Early warning commands can be generated based on the severity level of defects, and the early warning commands and defect information can be pushed to the host computer monitoring and early warning unit simultaneously.

[0150] The host computer monitoring and early warning unit enables the visualization of monitoring data, 3D imaging maps, and defect information. It also performs defect classification and early warning, historical data query, trend analysis, and remote system control. Specific functions include:

[0151] The built-in visualization display sub-unit uses 3D visualization technology to display the real-time 3D wave velocity field, permeability field, and 3D structural defect imaging map of the entire dam area. It can realize operations such as zooming in, zooming out, rotating, and segmenting the map to intuitively present the location, type, and scale of defects. At the same time, it can display real-time monitoring data, defect information, early warning status, etc., to realize one-stop display of monitoring information.

[0152] The built-in data analysis subunit can statistically analyze historical monitoring data, realize trend change analysis of dam structural parameters, predict the development trend of defects, and provide data support for dam safety assessment and operation and maintenance decisions. At the same time, it can automatically generate monitoring reports, the content of which can be customized according to needs, and support the export and printing of reports.

[0153] It has a built-in hierarchical early warning subunit that receives early warning instructions from the tomographic imaging processing unit and triggers corresponding audio-visual, pop-up, and other early warning prompts according to the severity level of the defect. At the same time, it can push early warning information to designated mobile terminals and management platforms to achieve multi-terminal synchronization of early warning information.

[0154] The built-in remote control subunit can send remote control commands to the controllable low-frequency excitation module, distributed sensor acquisition module, tomographic imaging processing unit, etc., to realize remote operations such as monitoring cycle adjustment, excitation parameter adjustment, sensor unit start and stop. At the same time, it can monitor the working status of each module / unit in real time, realizing remote operation and maintenance and fault diagnosis of the system.

[0155] It has system configuration and expansion functions, and can flexibly configure the system's monitoring parameters, abnormal thresholds, early warning rules, etc. according to the monitoring needs of the dam. It also supports the expansion and upgrading of system modules.

[0156] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for monitoring the safety of a dam, characterized in that, The method includes the following specific steps: S1: In the early stage of dam completion or under structurally healthy conditions, the original response data of the dam body and foundation are obtained through active excitation and distributed acquisition, and the initial three-dimensional wave velocity reference field and three-dimensional permeability reference field are inverted and constructed. S2: Multiple controllable low-frequency elastic wave excitation sources are arranged in preset spatial locations along the underwater area of ​​the dam's upstream side, the dam crest, the gallery, the curtain grouting area of ​​the dam foundation, and the dam shoulders on both banks. S3: Deploy distributed sensor arrays inside the dam body, seepage prevention curtain, dam foundation, drainage gallery and rock foundation on both banks. Under the unified clock synchronization, each sensor unit collects the original vibration waveform data of the excitation wave propagating in the medium in real time, and extracts the vibration response, propagation delay, amplitude attenuation, phase shift and energy distribution data in the propagation path. S4: Based on the elastic wave equation and the seepage field equation, a dual-field coupled forward model is constructed. The objective function is optimized by a regularized iterative inversion algorithm. After iterative convergence, the real-time three-dimensional wave velocity field and three-dimensional permeability field of the dam body and foundation are reconstructed. S5: Compare the real-time three-dimensional wave velocity field, permeability field and reference field element by element, calculate the relative change rate of wave velocity and permeability, identify the defect type based on the preset threshold and abnormal combination features, and generate a three-dimensional structural defect imaging map through volume rendering technology. S6: The ray tracing positioning algorithm combined with the three-dimensional mesh element coordinates is used to complete the three-dimensional precise positioning of defects. The severity level of defects is divided according to the abnormal amplitude, and the graded early warning is triggered and the positioning, type and level information are output. S7: Locate, quantitatively assess, and classify defects for early warning, and execute the process cyclically to achieve long-term online monitoring.

2. The dam safety monitoring method according to claim 1, characterized in that, In step S1, during the initial stage of dam completion or when the structure is in a healthy state, a three-dimensional distributed sensing array is used to collect the original data of the time-domain response, propagation delay, amplitude attenuation, and phase shift of the excitation wave propagating in the dam body, dam foundation, and anti-seepage curtain medium under high-precision clock synchronization. After preprocessing the collected data by noise reduction filtering, effective signal acquisition, and amplitude and phase correction, the elastic wave equation is used as the forward calculation model of the wave field, and the continuity and motion equation of the seepage field are used as parameter constraints. The three-dimensional space of the dam body and foundation is finely meshed, and the iterative inversion algorithm is used to calculate the medium wave velocity and permeability distribution parameters on a volume element basis. Through multiple rounds of excitation acquisition and iterative optimization, measurement errors are eliminated, and finally, an initial three-dimensional wave velocity reference field and a three-dimensional permeability reference field covering the entire dam main structure and foundation are constructed.

3. The dam safety monitoring method according to claim 1, characterized in that, In step S4, the elastic wave equation serves as the basis for forward wavefield calculation. A dual-field coupled forward model is constructed by coupling the seepage field continuity equation and the motion equation. Real-time response data collected by a distributed sensor array is used as inversion constraints. A regularized iterative tomographic inversion algorithm is employed to reconstruct the parameter field. The inversion objective function is used as the core of iterative optimization. The baseline model parameter field under dam health conditions is used as the initial iteration value. The wave velocity and permeability model parameters are updated round by round using the gradient descent method. The objective function value is calculated after each iteration until the rate of change of the objective function is ≤10%. -6 The convergence conditions are met, and the real-time reconstruction of the three-dimensional wave velocity field and the three-dimensional permeability field is completed.

4. The dam safety monitoring method according to claim 3, characterized in that, In step S4, a regularized iterative tomographic inversion algorithm is used to reconstruct the parameter field in order to invert the objective function. As the core of iterative optimization, the baseline model parameter field under the dam's healthy state is used. Using the initial iteration values, the wave velocity and permeability model parameters are updated round by round using the gradient descent method. The objective function value is calculated after each iteration until the rate of change of the objective function is ≤10. -6 The convergence condition is satisfied, where, It is the inversion objective function, used to measure the deviation between the model's calculated values ​​and the measured values, as well as the degree of model constraint. It is the model parameter field to be inverted, which includes wave velocity and permeability parameters in the three-dimensional space of the dam body and foundation; It is the observation data weight matrix, used to balance the observation weights of response data from different sensor channels and different locations; It is the excitation response data obtained from the actual measurement of the distributed sensing array, which includes information on wave propagation delay, amplitude, phase and energy attenuation; It is a forward model operator for elastic wave-seepage field coupling, which characterizes the mapping relationship between model parameters and excitation response data; It is a regularization coefficient used to suppress inappropriate issues during the inversion process. It is the model parameter weight matrix, used to constrain the spatial smoothness and physical rationality of the wave velocity field and the permeability field; It is the baseline model parameter field under the healthy state of the dam, namely the initial three-dimensional wave velocity baseline field and permeability baseline field.

5. A dam safety monitoring method according to claim 1, characterized in that, In step S5, the real-time three-dimensional wave velocity field and three-dimensional permeability field are compared with the initial three-dimensional wave velocity reference field and three-dimensional permeability reference field under the dam's healthy state, one by one, on a volumetric basis. The real-time wave velocity, real-time permeability, and corresponding reference values ​​of each three-dimensional grid volumetric element are extracted. The abnormal amplitude of each volumetric element is calculated using the wave velocity relative change rate formula and the permeability relative change rate formula. A unified and adaptively fine-tunable anomaly detection threshold is set, wherein the basic anomaly threshold is set as: wave velocity relative change rate Relative change rate of penetration 3D mesh elements that meet this condition are classified as structurally abnormal elements, while elements that do not meet this threshold are classified as healthy elements. Based on this, according to and By combining the abnormal amplitude combination characteristics with the differences in the physical properties of dam defects, the type of structural defect can be accurately determined: if and It was determined to be a tiny leakage channel; if Between 10% and 20% and If the area is between 10% and 30%, it is considered an internal loose zone; if the local volume element... Surrounding elements No obvious abnormalities and If it is determined to be a crack or void defect, and Then it is determined to be a healthy area.

6. A dam safety monitoring method according to claim 5, characterized in that, In step S5, the abnormal amplitude of each volume element is calculated using the formulas for the relative rate of change of wave velocity and the relative rate of change of permeability. The formula for the relative rate of change of wave velocity is: The formula for the relative rate of change of penetration is: In the formula, The relative rate of change of wave velocity is used to determine the degree of abnormality in the compactness of the dam medium and the structural integrity. The medium wave velocity value obtained through real-time inversion; This is the reference value for the medium wave velocity under healthy conditions; It is the relative rate of change of permeability, used to determine the seepage characteristics of the dam body and the degree of abnormality of the seepage channels; It is the medium permeability value obtained by real-time inversion; It is the benchmark value of media permeability under healthy conditions.

7. A dam safety monitoring method according to claim 1, characterized in that, In step S6, a ray tracing positioning algorithm combined with the spatial coordinate information of three-dimensional mesh elements is used to accurately locate the marked structural anomaly area in three-dimensional space. First, a global geodetic coordinate system for the dam is established, and the spatial coordinates of the three-dimensional mesh elements are precisely calibrated with the actual structural position of the dam. Taking each controllable low-frequency excitation source as the ray emission starting point and the corresponding distributed sensing unit receiving the excitation response signal as the ray reception endpoint, the propagation path of each ray in the dam body and foundation medium is calculated by combining the three-dimensional wave velocity field parameters obtained by real-time inversion, and the refraction and reflection errors of elastic waves in different media are corrected. Ray coverage analysis is performed on the marked anomaly elements, and all ray beams penetrating the anomaly element are counted. By fitting the intersection coordinates of multiple ray beams and spatial interpolation calculation, the accurate three-dimensional spatial coordinates of the anomaly element are determined, eliminating the deviation of single ray positioning. For continuously distributed anomaly elements, the spatial extension trajectory, length, and cross-sectional dimensions of the defect are fitted by the continuity analysis of the ray tracing path. For discrete and abruptly changing anomaly elements, the center point coordinates and influence range of the defect are determined by focusing the intersection points of the ray beams, and finally, the three-dimensional spatial positioning of the defect is achieved.

8. A dam safety monitoring method according to claim 1, characterized in that, In S6, based on the relative rate of change of wave velocity Relative change rate of permeability Based on the abnormal amplitude, combined with the defect's spatial range and extension length, the severity of the defect is divided into four levels: Level I: Between 5% and 10% Between 10% and 20%, the defect scale is small and there is no obvious development trend; Level II: Between 10% and 20% Between 20% and 50%, the defects have reached a certain scale and require regular monitoring. Level III: Between 20% and 30%, Between 50% and 80%, defects show a growing trend and are likely to cause safety hazards; Level IV: , The defect has affected the structural stability and needs to be addressed immediately.

9. A dam safety monitoring method according to claim 1, characterized in that, In step S6, the determination results of defect type and abnormality degree are associated with each volume element in the reconstructed three-dimensional wave velocity field and three-dimensional permeability field. Attribute annotation is performed on each three-dimensional mesh volume element. Combined with volume rendering technology, different color gradients and textures are used to distinguish healthy areas, various defects and different abnormalities, ultimately forming a high-resolution three-dimensional real-time parameter field.

10. A dam safety monitoring system, applicable to the dam safety monitoring method according to any one of claims 1-9, characterized in that, The system includes: Controllable low-frequency excitation module: performs time-sequence encoding and time-division transmission of controllable low-frequency elastic wave excitation signal, adaptively adjusts relevant signal parameters, and cooperates with the calibration signal of the transmission system to support the establishment of the reference field; Distributed sensing acquisition module: performs multi-channel synchronous acquisition and preprocessing of excitation wave response data, providing data support for dual-field coupling inversion and ray tracing positioning; Clock synchronization module: It adopts a dual-mode clock synchronization strategy to provide a high-precision unified time reference and ensure the time consistency of excitation transmission, data acquisition and positioning calculation; Data transmission module: It adopts a composite transmission method to adapt to different monitoring scenarios, realizes data transmission, and has data caching and encryption functions; Tomographic imaging processing unit: integrates data processing, dual-field coupling inversion, three-dimensional imaging, defect identification, ray tracing positioning and data storage functions, and completes the core processing of monitoring data and defect judgment; Upper computer monitoring and early warning unit: realizes the visualization of monitoring data, three-dimensional imaging map and defect information, supports historical data query, trend analysis and automatic report generation, and completes defect classification early warning and remote system control.

Images