A method for evaluating the flexibility of a breakwater structure and predicting the damage state probability

Abstract

This invention relates to the field of port and coastal engineering safety assessment technology, and discloses a method for evaluating the toughness of guide seawall structures and predicting the probability of damage states. The method includes: collecting and preprocessing multi-source heterogeneous monitoring data; constructing a time-varying Copula damage mode coupling model based on the monitoring sequence, and calculating the correlation coefficient matrix between failure modes; calculating the length of the damage space correlation feature and dynamically generating generalized evaluation units; introducing a historical cumulative dissipated energy index to construct a three-dimensional dynamic vulnerability model, simulating the power-law degradation of structural resistance with cumulative damage, and predicting the instantaneous probability distribution of each damage level. Finally, the system functional potential energy is calculated, and emergency repair resources are converted into a generalized repair potential energy flow, outputting a comprehensive toughness index through weighted integration. This invention comprehensively considers the time-varying coupling characteristics of failure modes and the cumulative damage effect, and incorporates emergency resource scheduling into the evaluation system, achieving dynamic and accurate quantification of the toughness of guide seawall structures.

Description

Technical Field

[0001] This invention relates to the field of port and coastal engineering safety assessment technology, specifically a method for evaluating the toughness of a breakwater structure and predicting the probability of damage status. Background Technology

[0002] As a crucial component of port and coastal protection engineering, guide seawalls endure the continuous dynamic loads of wind, waves, and tidal cycles. Their structural safety directly impacts port operational efficiency and the stability of the protected waters. Existing methods for assessing the structural safety of guide seawalls primarily rely on deterministic analysis based on design specifications or traditional static reliability theory, focusing on evaluating the structure's ability to withstand extreme loads at a single moment.

[0003] However, during actual service, guide dike structures often face complex environmental erosion and multi-physics interactions. Existing assessment techniques typically assume that multiple failure modes, such as revetment block instability, toe scouring, and dike slippage, are independent of each other, or only consider their static correlations, neglecting the time-varying coupling characteristics of these failure modes under different load levels and damage stages. This leads to biases in the prediction of the overall system failure risk. Furthermore, existing models often use fixed evaluation units for vulnerability analysis, lacking dynamic definition of the spatial correlation evolution range caused by local damage, and failing to fully quantify the impact of cumulative dissipated energy due to hysteretic deformation on resistance attenuation throughout the structure's life cycle. Consequently, they fail to accurately reflect the physical degradation patterns of the structure under service conditions.

[0004] Meanwhile, in terms of resilience assessment, current research mostly focuses on the remaining functional capacity after structural damage, that is, on the decline phase of the performance curve, while the quantification of the post-disaster recovery phase often remains at the level of qualitative description or simple recovery time estimation. The existing assessment system has not yet been able to effectively physically map and integrate the emergency repair resource reserves and mobilization rate in the physical world with the recovery dynamics of the structural system, making it difficult for the assessment results to directly support dynamic maintenance resource scheduling and decision optimization. Summary of the Invention

[0005] In view of the prior art, the present invention provides a method for evaluating the toughness of guide embankment structures and predicting the probability of damage states, aiming to solve the problems in the prior art where the safety assessment of guide embankments does not fully consider the time-varying coupling characteristics of multiple failure modes, the spatial correlation evolution of damage, and the impact of cumulative damage on the attenuation of structural resistance.

[0006] A method for evaluating the toughness of a guide embankment structure and predicting its damage state probability includes the following steps: S1. Collect multi-source heterogeneous monitoring data, clean and synchronize the data with time, and construct a standardized state monitoring vector sequence.

[0007] S2. Based on the standardized state monitoring vector sequence, construct a damage mode coupling model based on the time-varying Copula function, and output the correlation coefficient matrix that evolves over time.

[0008] S3. Obtain the maximum correlation coefficient modulus in the correlation coefficient matrix, calculate the length of the damage space related features based on the change of the maximum correlation coefficient modulus, and divide the generalized evaluation units accordingly.

[0009] S4. Within the generalized evaluation unit, a three-dimensional dynamic vulnerability model considering accumulated energy dissipation is constructed, and the instantaneous probability distribution vector of the guide embankment structure at each damage level is predicted based on the three-dimensional dynamic vulnerability model.

[0010] S5. Calculate the system functional potential energy based on the instantaneous probability distribution vector, and map the emergency repair resources into a generalized repair potential energy flow. By performing a weighted integral on the system functional potential energy and the generalized repair potential energy flow, output a comprehensive resilience index for quantitative evaluation criteria and decision-making basis.

[0011] Furthermore, the vector elements in the standardized state monitoring vector sequence include key indicators of environmental load and structural response, specifically significant wave height, real-time tide level, maximum scour depth at the toe of the dike, cumulative horizontal displacement of characteristic points on the dike body, and pore water pressure. The preprocessing process employs a Kalman filter algorithm to denoise the original time series, eliminating environmental interference, and uses linear interpolation to resample heterogeneous data collected by different sensors to a unified time step, ensuring the spatiotemporal synchronization of multi-source data.

[0012] Furthermore, in constructing the damage mode coupling model based on the time-varying Copula function, the main failure modes of the guide embankment structure are identified, including face block instability, toe scouring, and embankment slippage, and the edge cumulative distribution of each failure mode is fitted. A time-varying Copula function is used to describe the joint probability distribution among the failure modes. Historical monitoring data is extracted using a sliding time window technique, and the posterior distribution of the time-varying parameter vector describing the correlation structure between variables is updated in real time using a Bayesian inference algorithm, thereby calculating the correlation coefficient matrix reflecting the coupling strength between each failure mode at the current moment.

[0013] Furthermore, the calculation of the length of the damage space-related features is based on the dynamic changes of the correlation coefficients. The maximum correlation coefficient magnitude in the correlation coefficient matrix is ​​extracted, which reflects the coupling strength between different failure modes at the current moment. The length of the damage space-related features is calculated using a preset nonlinear mapping function. This function is set such that the length of the damage space-related features increases nonlinearly between a preset base length and a maximum allowable evaluation length as the maximum correlation coefficient magnitude increases, thereby characterizing the diffusion characteristics of the damage space's influence range under high coupling conditions.

[0014] Furthermore, the generation process of the generalized evaluation unit includes: using the longitudinal axis of the guide embankment as a reference, determining whether the geometric center distance between adjacent initial discrete micro-elements is less than the calculated length of the damage space-related characteristics. Adjacent initial discrete micro-elements that satisfy the distance condition are meshed to generate a generalized evaluation unit, and the equivalent physical parameters of this unit are redefined, serving as the dynamic physical evaluation boundary for subsequent vulnerability analysis.

[0015] Furthermore, before constructing the three-dimensional dynamic vulnerability model, the historical cumulative dissipation energy index of the guide embankment structure is calculated. This index is defined as the ratio of the cumulative plastic deformation energy of the structure during its entire life cycle hysteresis cycle to the design ultimate energy storage capacity, and is used to quantify the cumulative damage effect of the structure.

[0016] Furthermore, the process of constructing a three-dimensional dynamic vulnerability model considering cumulative dissipated energy includes: introducing the predicted load intensity at the next moment and the historical cumulative dissipated energy index as independent variables into the vulnerability function, thereby constructing a three-dimensional dynamic vulnerability model that includes the predicted load intensity, the historical cumulative dissipated energy index, and the damage state. The vulnerability function in the three-dimensional dynamic vulnerability model sets the median resistance to decrease power-lawfully with the increase of the historical cumulative dissipated energy index, thus describing the physical law of structural resistance degradation due to cumulative damage.

[0017] Furthermore, when predicting the instantaneous probability distribution vector of the guide embankment structure at each damage level, a non-homogeneous state transition matrix is ​​generated based on the three-dimensional dynamic vulnerability model. This matrix is ​​then used to calculate the instantaneous probability distribution vector of the guide embankment structure at each damage level (intact, slightly damaged, moderately damaged, and severely damaged) at the next moment.

[0018] Furthermore, the system functional potential energy is defined as the weighted sum of the probabilities of each damage state and their corresponding functional retention coefficients. When mapping emergency repair resources to a generalized repair potential energy flow, based on preset resource conversion efficiency, rate saturation factor, and damage state threshold for initiating repair, a functional relationship is established between the emergency repair resource reserve and mobilization rate and the system recovery capability. Material resources are quantified as a generalized repair potential energy flow that performs work on the damaged system, thereby characterizing the driving force of external intervention on system recovery.

[0019] Furthermore, the specific calculation method for the output comprehensive resilience index is as follows: within a predetermined evaluation period, the product of the system functional potential energy and the functional maintenance weight coefficient, and the product of the generalized repair potential energy flow and the repair resource weight coefficient are weighted and summed, and the summation result is integrally applied over the time domain of the evaluation period to obtain the comprehensive resilience index.

[0020] This invention provides a method for evaluating the toughness of guide embankment structures and predicting the probability of damage states. It has the following beneficial effects: 1. This invention constructs a damage mode coupling model based on a time-varying Copula function, which can update the correlation coefficients between various failure modes in real time. This effectively solves the problem that traditional assessment methods ignore the time-varying coupling characteristics between multiple failure modes. By using Bayesian inference to correct the coupling strength in real time, the safety assessment of the guide seawall structure in complex marine environments is more in line with actual working conditions. This avoids assessment bias caused by assuming that the failure modes are independent of each other, and improves the accuracy of predicting the failure probability of the structural system.

[0021] 2. This invention introduces the damage space-related characteristic length and cumulative dissipation energy index to achieve dynamic quantification of damage space evolution and structural resistance decay. By dynamically adjusting the size of the generalized evaluation unit according to the coupling strength and constructing a three-dimensional dynamic vulnerability model that considers historical cumulative damage, this invention can simulate the physical law of power-law degradation of structural resistance with service time and cumulative deformation, thereby overcoming the shortcomings of existing technologies that fail to fully consider the impact of cumulative damage on structural safety and improving the reliability of damage state prediction throughout the entire life cycle of the guide embankment.

[0022] 3. This invention establishes a resilience quantification system that maps emergency repair resources to a generalized repair potential energy flow, achieving a unified assessment of engineering structural performance and emergency management resources. By converting the physical world's resource reserves and mobilization rate into energy indicators that do work on the system, and weighting and integrating them with the system's functional potential energy, this invention not only assesses the robustness of the structure itself, but also quantifies the driving force of external intervention on system recovery, providing a scientific quantitative basis for dike maintenance decisions and emergency resource scheduling. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating the method for evaluating the toughness of a guide embankment structure and predicting the probability of damage state according to the present invention. Figure 2 This is a functional module architecture diagram of the computing processing system in this invention.

[0024] Among them, 10 is the monitoring data processing module; 20 is the damage mode coupling modeling module; 30 is the dynamic evaluation domain reconstruction module; 40 is the damage state probability prediction module; and 50 is the toughness quantification evaluation module. Detailed Implementation

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] See attached document Figure 1 This embodiment provides a method for evaluating the toughness of a guide embankment structure and predicting the probability of damage state, which is executed by a computational processing system. The computational processing system includes a monitoring data processing module 10, a damage mode coupling modeling module 20, a dynamic evaluation domain reconstruction module 30, a damage state probability prediction module 40, and a toughness quantification evaluation module 50. Specifically, it includes the following steps: Step S100 involves using the monitoring data processing module 10 to acquire and vectorize heterogeneous monitoring data from multiple sources. Step S100 uses wave buoys, tide gauges, multibeam sonar, and embedded sensors deployed at the guide dike site to acquire environmental load variables and structural response variables under the service environment of the guide dike. The monitoring data processing module 10 performs denoising, outlier removal, and time step synchronization operations on the acquired raw time series signals to construct a standardized state monitoring vector sequence containing significant wave height, real-time tide level, scour depth, horizontal displacement, and pore water pressure.

[0027] Step S200: A damage mode coupling model based on a time-varying Copula function is constructed using the damage mode coupling modeling module 20. Based on the state monitoring vector sequence output in step S100, step S200 identifies the edge distribution characteristics of the main failure modes of the guide embankment. These failure modes include revetment block instability, toe scouring, and embankment slippage. The damage mode coupling modeling module 20 uses a time-varying Copula function to describe the joint probability distribution among these failure modes and applies a sliding time window technique combined with a Bayesian inference algorithm to update the parameters of the Copula function in real time, outputting the correlation coefficient matrix of each failure mode evolving over time.

[0028] Step S300: Dynamic evaluation domain reconstruction based on damage correlation drift is performed using the dynamic evaluation domain reconstruction module 30. Step S300 receives the correlation coefficient matrix output in step S200 and calculates the length of the damage space correlation feature based on the numerical changes of the correlation coefficients. The dynamic evaluation domain reconstruction module 30 dynamically adjusts the meshing strategy of the guide embankment physical evaluation model according to the length of the damage space correlation feature, merging adjacent initial discrete micro-elements that satisfy the correlation conditions to generate generalized evaluation units, thereby determining the physical evaluation boundary for vulnerability analysis at the current moment.

[0029] Step S400: The damage state probability prediction module 40 constructs a three-dimensional dynamic vulnerability model considering accumulated dissipated energy and predicts the state transition probability. Within the generalized evaluation unit determined in step S300, the historical accumulated dissipated energy index of the structure is calculated, and this index is introduced as an independent variable into the vulnerability function to construct a three-dimensional dynamic vulnerability surface encompassing load intensity, accumulated dissipated energy, and damage state. Based on this surface, the damage state probability prediction module 40 generates a non-homogeneous state transition matrix and calculates the instantaneous probability distribution vector of the guide embankment structure at each damage level at the next prediction time.

[0030] Step S500 involves quantifying the resilience index based on the generalized repair potential energy field using the resilience quantification evaluation module 50. Step S500 calculates the system functional potential energy of the guide embankment based on the instantaneous probability distribution vector of the damage state output in step S400. Simultaneously, the resilience quantification evaluation module 50 physically maps the preset emergency repair resource reserves and mobilization rate into a generalized repair potential energy flow that performs work on the damaged system. The resilience quantification evaluation module outputs a comprehensive resilience index characterizing the full life-cycle performance of the guide embankment structure by calculating the weighted integral of the system functional potential energy and the generalized repair potential energy flow over a predetermined time domain.

[0031] See attached document Figure 2 The computational processing system mainly includes a monitoring data processing module 10, a damage mode coupling modeling module 20, a dynamic evaluation domain reconstruction module 30, a damage state probability prediction module 40, and a toughness quantification evaluation module 50. These modules are connected via a data communication interface to achieve dynamic evaluation of the guide embankment structure.

[0032] The monitoring data processing module 10 is used to collect and standardize data on the service environment and structural response of the guide dike. The specific execution process of this module includes step S101, receiving data streams from field sensing devices via a hardware interface. These field sensing devices include: wave buoys deployed at the leading edge of the guide dike to measure wave height and period; tide gauges installed in the harbor to acquire real-time tide levels; multibeam sonar mounted on inspection vessels or fixed supports to scan the mud surface elevation in the dike toe area; and fiber optic grating sensors and pore water pressure gauges embedded inside the dike body to monitor structural displacement and internal pore pressure. In step S102, the monitoring data processing module 10 maps the aforementioned heterogeneous data to a unified time frame. The state monitoring vector below : ,in, For the effective wave height, For real-time tide levels, The maximum scouring depth at the toe of the dike. The cumulative horizontal displacement of the characteristic points of the embankment. The value is the pore water pressure. In step S103, this module uses a Kalman filter algorithm to denoise the original time series and uses linear interpolation to align data from different sampling frequencies to a unified time step. Output a standardized sequence of state monitoring vectors.

[0033] Damage Mode Coupling Modeling Module 20 is used to analyze the nonlinear dependencies between multidimensional failure modes. In step S201, based on the mechanical properties of the guide embankment structure, this module identifies the instability of the revetment blocks, toe scouring, and overall slippage of the embankment as the main failure modes, and fits the edge cumulative distribution of each failure mode using an extreme value distribution function. In step S202, this module establishes a multidimensional joint distribution model: ,in, For Copula functions, This module generates a time-varying parameter vector describing the correlation structure between variables. In step S203, the module sets a sliding time window, extracts historical monitoring data within the current window, and uses Bayesian inference to update the time-varying parameter vector describing the correlation structure between variables. The posterior distribution is then used to calculate the correlation coefficient matrix between each failure mode at the current time. The selection of the Copula function and the specific numerical solution of Bayesian inference are standard techniques in the field of statistics and will not be elaborated upon here.

[0034] The dynamic evaluation domain reconstruction module 30 is used to adjust the spatial topology of the physical evaluation model based on statistical correlation. This module extracts the correlation coefficient matrix in step S301. The maximum correlation coefficient magnitude in This characterizes the coupling strength between different failure modes at the current moment. In step S302, the length of the relevant feature in the damage space is calculated using a preset nonlinear mapping function. : ,in, The foundation length of a single caisson section. To be the maximum allowed evaluation length, For shape parameters, This is the correlation threshold. In step S303, the module uses the longitudinal axis of the guide embankment as a reference and sets the geometric center distance to less than [a certain value]. Adjacent initial discrete elements are merged into a mesh to generate a new generalized evaluation unit. And redefine the equivalent physical parameters of the unit, using them as the physical object for subsequent probabilistic analysis.

[0035] The damage state probability prediction module 40 is used to predict the evolution of the structural damage state within the reconstructed evaluation domain. In step S401, this module calculates the historical cumulative dissipation energy index of the structure. This index is defined as the ratio of the cumulative plastic deformation energy of the structure during hysteresis cycles to the design ultimate energy storage capacity. In step S402, a three-dimensional dynamic vulnerability function incorporating the cumulative dissipation energy index is constructed: ,in, To predict the load intensity at the next moment, The standard normal cumulative distribution function is... The total logarithm standard deviation, The median of the resistance decreases with the cumulative energy dissipation exponent, representing the structural resistance as a function of the cumulative energy dissipation exponent. The value decreases power-lawfully as the vulnerability increases. In step S403, the module generates a non-homogeneous state transition matrix based on the vulnerability function and calculates the instantaneous probability distribution vectors of the guide embankment structure at each level of integrity, minor damage, moderate damage, and severe damage in the next moment. .

[0036] The toughness quantification evaluation module 50 is used to comprehensively calculate the toughness index of the guide embankment. In step S501, this module calculates the toughness index based on the instantaneous probability distribution vector. Computational system functional potential This is the weighted sum of the probabilities of each damage state and their corresponding functional retention coefficients. In step S502, the input emergency repair resource reserve is... With mobilization rate Converted into generalized repair potential flow : ,in, To improve resource conversion efficiency, As the rate saturation factor, It is a step function. The primary damage level at the current moment. The damage state threshold for initiating repair is set. In step S503, the module sets a predetermined evaluation period. The system's functional potential energy and generalized repair potential energy flow are weighted and integrated to output a comprehensive resilience index. The calculation formula is as follows: ,in, To maintain the weighting coefficients for functionality, To repair the resource weighting coefficients. This comprehensive resilience index. This is used to quantify the ability of dikes to maintain their function and recover using external resources throughout a disaster.

[0037] This embodiment obtains raw physical data by constructing a distributed monitoring network. The specific steps include: Step S111: Collect environmental load data. Deploy wave monitoring equipment (buoys or ADCPs) on the deep-water side at a distance of 2 to 3 times the design wavelength from the leading edge of the dike structure to obtain the effective wave height. And the spectral peak period; deploy water level gauges (radar or pressure type) on the inner side of the harbor or in relatively still waters to obtain real-time tide levels. .

[0038] Step S112: Detect the scouring status of the breakwater toe. A multibeam echo sounder system (mobile survey or fixed scan) is used to acquire a 3D point cloud of the seabed in the breakwater toe area. The current elevation data is compared with the reference seabed elevation to calculate the difference and extract the maximum scouring depth of the breakwater toe. .

[0039] Step S113: Monitor the displacement and deformation of the embankment structure. A monitoring subsystem is deployed at the settlement observation points on the top of the guide embankment breast wall or at the caisson joints: Fiber Bragg grating (FBG) sensors are used to detect strain and relative displacement, or GNSS receivers are used to acquire absolute three-dimensional coordinate changes, thereby calculating the cumulative horizontal displacement of embankment feature points perpendicular to the axis. .

[0040] Step S114: Monitor the pore water pressure inside the dike. Vibrating wire pore water pressure gauges, equipped with filter bags, are pre-embedded in the riprap foundation, inverted filter layer, and core stone area to collect the pore water pressure after temperature compensation correction. .

[0041] Data from all the aforementioned monitoring nodes are aggregated to the shore-based data center via underwater optical cables or wireless modules.

[0042] In this embodiment, a state monitoring vector is constructed. The process aims to standardize monitoring data into a uniform input format, and specifically includes the following steps: Step S121: Extract environmental load intensity parameters. The system converts the monitored marine environmental data into statistical parameters characterizing the intensity of external loads. For wave data, the intercept time is... The water surface elevation time series within the previous preset time period (e.g., 20 to 30 minutes) is used to calculate the significant wave height using the zero-point crossing method or the spectrum analysis method. This parameter represents the wave force energy level exerted on the structure under the current sea state. For tide data, the time is read directly. Real-time tide level This parameter determines the specific elevation area where waves act on the guide seawall revetment and the probability of wave overtopping. These two parameters together constitute the environmental load component in the state vector.

[0043] Step S122: Extract engineering requirement parameters. The system transforms structural response monitoring data into key indicators characterizing the physical state of the structure. For scour at the toe, based on the point cloud difference results from sonar scanning, the maximum vertical erosion depth is retrieved within a predetermined evaluation unit range and defined as the maximum scour depth at the toe. This indicator reflects the risk of local instability of the foundation. Regarding the deformation of the embankment, the calculation is performed at the current moment. Feature point coordinates relative to the initial time The absolute value of the projection difference of the coordinates in the direction perpendicular to the guide embankment axis is used as the cumulative horizontal displacement of the characteristic point of the embankment. This indicator reflects the overall slippage or overturning trend of the levee. For internal seepage, the pore water pressure at the measuring point in the levee core is directly extracted. This indicator reflects the effective stress state and liquefaction potential of the soil.

[0044] Step S123: Construct a multi-dimensional state monitoring vector. Combine the environmental load parameters extracted in step S121 with the structural response parameters extracted in step S122 in a predetermined order to construct the time-varying vector. Five-dimensional column vector : State monitoring vector A complete description of the guide embankment at time The coupling state of load and response. State monitoring vector. Each element in the vector corresponds to the same physical evaluation domain and the same discrete time node. If the sampling frequencies of the sensors are inconsistent, the sampling time of the sensor with the lowest frequency is used as the benchmark. Statistical resampling is performed on the high-frequency data, or linear interpolation is performed on the low-frequency data to ensure time synchronization within the vector. This vector serves as the basic data structure and is directly input into the subsequent damage mode coupling modeling module.

[0045] In this embodiment, data cleaning and synchronization processing of multi-source heterogeneous monitoring data is performed, specifically including the following steps: Step S131: Remove outliers from the monitoring data. For outliers caused by electromagnetic interference or communication failures in the sensors, the Raida criterion (3...) is used. The criteria are used for identification and elimination. For any monitoring variable sequence... Select the current time A sliding window centered on the data point is used to calculate the mean of the data within the window. with standard deviation If the monitored value meets the requirements... If the data point is an outlier, it is determined to be an outlier and replaced with the linear interpolation result of the adjacent normal data before and after that time.

[0046] Step S132: Smooth and denoise the monitoring data. This involves calculating the cumulative horizontal displacement of key points on the embankment. With pore water pressure For continuous variables containing random observation noise, the Kalman filter algorithm is used for optimal estimation. The state equation and observation equation are established as follows: ; ; in, For a moment The actual state value, For a moment Sensor observations Let this be the state transition matrix (here, we set it as the identity matrix, assuming that the physical quantities change steadily over a short period of time). For the observation matrix, and These are process noise and observation noise, respectively, both following a Gaussian distribution.

[0047] Update the state estimate using the following recursive formula: ; ; ;in, This is the filtered value after noise reduction. For the prediction error covariance matrix, To observe the noise covariance matrix, This is the Kalman gain. This step effectively separates the true structural response trend from the monitored signal.

[0048] Step S133: Synchronize heterogeneous data in the time domain. Set a unified time reference axis. ,in Set the system's preset analysis step size (e.g., 1 hour). Implement differentiated resampling strategies for data with different sampling characteristics: for significant wave height... With real-time tide level Linear interpolation is used to map non-uniform sampling points to the time reference axis. Above; for the maximum scouring depth at the toe of the dike Due to the long sonar detection cycle, a zero-order hold strategy is adopted, that is, before new detection data is updated, it is assumed that the current sweep depth remains unchanged from the previous detection time; for the horizontal cumulative displacement of the embankment feature points after filtering in step S132... With pore water pressure The extraction method is used to directly extract the closest time reference point. The filtered values ​​are then processed as follows. After the above processing, all components are fully aligned in the time dimension, forming a standardized state monitoring vector sequence. And store it in the time series database for subsequent modules to call.

[0049] This embodiment is based on the state monitoring vector obtained from the aforementioned steps. The typical failure modes of the guide embankment structure are mathematically defined, and the statistical distribution characteristics of each physical quantity are fitted. Specifically, the steps include: Step S211: Define the random variables corresponding to typical failure modes. Based on the mechanical properties of the guide embankment structure, map the physical components in the monitoring vector to random variables characterizing specific failure modes. For the instability failure mode of the retaining wall block, the effective wave height is selected. As the master random variable This variable determines the magnitude of the drag force and uplift force on the revetment blocks. For the scouring failure mode at the toe, the maximum scouring depth at the toe is selected. As the master random variable This variable characterizes the degree of support weakening at the front edge of the riprap foundation. For the slippage failure mode of the embankment, the cumulative horizontal displacement of characteristic points on the embankment is selected. As the master random variable This variable characterizes the cumulative plastic deformation state of the levee under the combined action of wave thrust and internal pore water pressure. In this step, if pore water pressure is present... Abnormal fluctuations can be treated as a fourth-dimensional independent random variable. This embodiment uses a three-dimensional variable as an example for illustration.

[0050] Step S212: Construct a marginal probability distribution model for the random variables. Since levee failure is usually triggered by extreme environmental loads or extreme states of cumulative damage, this embodiment uses the generalized extreme value distribution (GEV) for the aforementioned random variables. The tail features are fitted. The probability density function (PDF) of the generalized extreme value distribution is then used. The definition is as follows: ,in, For position parameters, For scale parameters ( , For shape parameters. This formula applies to... The domain of definition. Correspondingly, its cumulative distribution function (CDF). for: For the master random variable If its statistical characteristics are non-negatively skewed and without obvious extreme value cutoff, a log-normal distribution can also be used to describe it. The specific distribution type is determined by the Akaike Information Criterion (AIC) or the Bayesian Information Criterion (BIC).

[0051] Step S213: Perform parameter estimation and marginal distribution transformation. Using historical monitoring datasets or observation samples within a sliding time window, employ maximum likelihood estimation (MLE) to perform parameter estimation on the parameter set in the above formula. Solve by constructing the log-likelihood function. : ,in For the sample size, For the first The optimal parameter estimate is obtained by taking the partial derivative of the parameter and setting it to zero using a numerical iterative algorithm. After obtaining the parameter estimates, the current time... Measured values ​​of each random variable Substitute into the cumulative distribution function Calculate the marginal cumulative probability value within the range [0,1]. : The cumulative probability value at the edge The dimensional differences of the original physical quantities are eliminated, and the standardized marginal distribution information is directly input into the subsequent time-varying Copula function to construct the joint distribution. The specific numerical calculation process for maximum likelihood estimation can be implemented using conventional algorithms such as the Newton-Raphson method by those skilled in the art, and will not be elaborated upon here.

[0052] This embodiment obtains the edge cumulative probability values ​​of each failure mode. Then, the dynamic dependency structure between significant wave height, maximum scour depth at the toe, and cumulative horizontal displacement of characteristic points on the dike body is described using a time-varying Copula function, specifically including the following steps: Step S221: Construct a joint cumulative distribution function model based on Sklar's theorem. According to Sklar's theorem, the joint cumulative distribution function of multidimensional random variables... It can be uniquely determined by its marginal distribution function and Copula function. This embodiment defines it at time [time]. Below, three main control random variables The time-varying joint cumulative distribution function is: ,in, These correspond to the significant wave heights. Maximum scouring depth at the toe of the dike Horizontal cumulative displacement of characteristic points of the embankment The edge cumulative probability values ​​are all in the range of [0,1]. For the selected Copula function, The correlation parameter matrix represents the time-varying characteristics of the coupling degree between different failure modes.

[0053] Step S222: Define the high-dimensional Gaussian Copula function structure. Based on the multidimensionality of the guide failure modes and the symmetry of the correlations between variables, this embodiment selects the Gaussian Copula as the connection function. Its mathematical expression is defined as: ,in, The inverse cumulative distribution function of the standard normal distribution is used to convert marginal probability values. Mapped to the standard normal space. For the correlation parameter matrix The parameter is the ternary standard normal joint cumulative distribution function. This function structure decouples the marginal distribution characteristics of each physical quantity from the correlation structure between variables, making the correlation analysis independent of the statistical properties of individual variables. Step S223: Derive the time-varying joint probability density function. Take the partial derivative of the joint cumulative distribution function to obtain the time-varying joint probability density function. : ,in, Let be the marginal probability density function of each random variable. Let be the Gaussian Copula density function, and its analytical expression is: ;in, The marginal probability vector, The determinant of the correlation parameter matrix. It is a third-order identity matrix. This is the inverse of the correlation parameter matrix. (Vector) Defined as the transformed standard normal variable vector: ; Step S224: Define the time-varying structure of the correlation parameter matrix. (Correlation parameter matrix) It is a symmetric positive definite matrix, whose diagonal elements are always 1, and whose off-diagonal elements are 1. Indicates the first The failure mode and the first Each failure mode at time [time] Linear correlation coefficient: ;in, This matrix is ​​updated through subsequent Bayesian inference or sliding window statistics to reflect the dynamic changes in the correlation between failure modes under extreme sea conditions (such as the increased correlation between face instability and breakwater slippage caused by increased wave height).

[0054] This embodiment utilizes monitoring data within a sliding time window and updates the correlation parameters in the Copula function in real time through Bayesian inference to capture non-stationary changes in the coupling relationships between failure modes. The process specifically includes the following steps: Step S231: Define the vector of parameters to be estimated and the sample set. Then, use the correlation parameter matrix... The independent off-diagonal elements are extracted to form the vector of parameters to be estimated. : ;in, These are the correlation coefficients between significant wave height and maximum scour depth at the toe, significant wave height and cumulative horizontal displacement of characteristic points on the dike body, and maximum scour depth at the toe and cumulative horizontal displacement of characteristic points on the dike body, respectively. Meanwhile, a length is set as... Using a sliding time window, the cumulative probability value sequence of the edges after edge distribution transformation within the current window is extracted as the observation sample set. : ;in, For a moment The marginal probability vector.

[0055] Step S232: Construct the posterior probability density function of the parameters. Based on Bayes' theorem, the parameter vector... In a given set of observation samples The posterior distribution under Proportional to the product of the likelihood function and the prior distribution: ;in, The prior distribution is denoted as [-1, 1]. Considering that the correlation coefficient ranges from [-1, 1], this embodiment uses a truncated normal distribution or a uniform distribution as the prior distribution. Let be the likelihood function of the sample set, which is obtained by multiplying the Copula density functions derived in S223: ;in, Let be the Gaussian Copula density function. For vector The reconstructed correlation parameter matrix.

[0056] Step S233: Sampling is performed using the Markov Chain Monte Carlo (MCMC) algorithm. Due to the complexity of the aforementioned posterior distribution analytical expression, this embodiment employs the Metropolis-Hastings algorithm for numerical solution. The specific process is as follows: In the In the next iteration, based on the current parameter state and suggested distribution (e.g., multidimensional normal random walk) to generate candidate samples .

[0057] Calculate the probability of acceptance : ; Generate uniformly distributed random numbers ,like Then accept the candidate sample and let Otherwise, refuse, and .

[0058] Repeat the above process until a predetermined number (e.g., 10,000) of convergent sample chains are obtained.

[0059] Step S234: Calculate the point estimates of the parameters. Remove the burn-in samples from the initial MCMC sampling period and use the mean of the remaining samples as the current time value. Time-varying parameter vectors describing the correlation structure between variables The estimated value: ,in, The number of valid samples, This is a posterior sample. Based on Update the correlation parameter matrix This matrix will be used for subsequent joint probability calculations and resilience evaluation. The specific programming implementation and convergence diagnosis of the MCMC algorithm are well-known techniques in the field of statistical computing and will not be elaborated upon here.

[0060] This embodiment constructs a discrimination mechanism based on damage correlation drift characteristics to monitor in real time whether the coupling state between failure modes has drifted and changed the overall toughness evaluation benchmark. Specifically, it includes the following steps: Step S311: Calculate the correlation drift rate index. This is done using the estimated values ​​of the time-varying parameter vector describing the correlation structure between variables, output from step S234. Calculate the current time. Compared to the previous moment correlation drift rate Its definition adopts the Euclidean norm form: ,in, The time step for relevance updates, For the set of correlation coefficient subscripts, This represents the L2 norm of the vector. The correlation drift rate. This characterizes the rate at which the coupling tightness between different failure modes changes under environmental loads or cumulative structural damage. For example, when the maximum scour depth at the toe of the dike... The rapid increase leads to the cumulative horizontal displacement of the embankment's characteristic points. This index will show a significant peak when the sensitivity to wave loads suddenly increases.

[0061] Step S312: Calculate the structural damage activation factor. To avoid false triggering due to correlation value fluctuations under low stress levels or minor environmental disturbances, structural physical state constraints are introduced. Extract the current cumulative horizontal displacement of the embankment feature points. Maximum scouring depth at the toe of the dike Constructing structural damage activators This factor is calculated using a weighted normalization formula: ,in, The allowable horizontal displacement limit value for the guide embankment design. To determine the maximum allowable brush depth limit, and Weighting coefficients (satisfying) This is used to balance the importance of the two damage modes in the triggering mechanism.

[0062] Step S313: Generate the evaluation domain reconstruction trigger signal. Construct a discrimination logic based on dual thresholds to generate a binary trigger signal. Set the drift rate threshold. With damage activation threshold The judgment logic is as follows: ; when When the signal is positive, it indicates a significant shift in the coupling relationships between the various failure modes of the guide embankment, rendering the original static evaluation domain inaccurate in describing the system's safety boundary. The system then initiates a subsequent dynamic evaluation domain reconstruction process. Conversely, if the signal is 0, the current evaluation domain setting is maintained, and routine monitoring continues. (Regarding thresholds...) and The numerical setting can be determined by those skilled in the art based on the structural importance level of the guide dike and the statistical distribution characteristics of historical monitoring data (such as taking the 95th percentile of the historical fluctuation range).

[0063] This embodiment establishes a quantitative mapping relationship between the statistical correlation between failure modes and the physical spatial evaluation range of the guide embankment after the evaluation domain reconstruction mechanism is triggered. The spatial correlation feature length is calculated using this algorithm. The minimum longitudinal physical length required for structural toughness analysis at this moment is defined to ensure that the evaluation domain can cover the spatial chain effects caused by damage coupling. This includes the following steps: Step S321: Calculate the global coupling strength index. This is based on the estimated value of the time-varying parameter vector describing the correlation structure between variables, output from step S234. Extracting elements Construct the global coupling strength that characterizes the overall correlation of the system at the current moment. Considering that different combinations of failure modes have different weights on the overall structural stability, a weighted sum of square roots is used for definition: ;in, This is the set of subscripts for correlation coefficients. The weight coefficients corresponding to the relevant relationships satisfy the following conditions: For example, for gravity-type guide dikes, the correlation coefficient between the maximum scour depth at the dike toe and the cumulative horizontal displacement of characteristic points on the dike body is... They are typically assigned a large weight because their coupling directly corresponds to the overall risk of capsizing. This global coupling strength... The larger the value, the stronger the mutual constraint and transmission effect between the failure modes, and the easier it is for local damage to a single section to trigger a chain reaction along the longitudinal axis of the guide embankment.

[0064] Step S322: Establish the mapping relationship between global coupling strength and the length of spatially relevant features in the continuous domain. Based on the physical mechanism that stronger correlation leads to a larger spatial influence range, this embodiment uses an exponential diffusion model to calculate the length of spatially relevant features in the continuous domain. : ;in, The lower limit of the evaluation domain length is usually taken as the width of a single caisson structure segment or the length of a single construction segment; To evaluate the upper limit of the domain length, the spatial correlation scale of the guide dike structure under hydrodynamic action is usually taken (such as one wavelength of the wave or the spatial autocorrelation distance of the foundation soil). The correlation sensitivity coefficient ( The formula () is used to adjust the rate at which the evaluation domain grows with increasing coupling strength. This formula ensures that the evaluation domain converges to local structural units under low correlation, while rapidly expanding to the macroscopic scale under high coupling.

[0065] Step S323: Determine the length of the discretized spatially relevant features. Considering that guide embankment projects are typically composed of discrete precast caissons or segmented cast-in-place blocks, the continuous values ​​calculated in step S322 are used... Mapped to physically feasible spatially relevant feature length Let the width of the basic structural unit of the guide dike be... (For example, the width of the caisson), use the rounding up strategy: ;In the formula, This indicates the floor function. The result determined through this step... This refers to the physical length of the reconstructed dynamic evaluation domain. This length determines the range of longitudinal measurement points and the number of structural units along the embankment that need to be included in the joint probability calculation in the subsequent resilience assessment model.

[0066] This embodiment is based on the calculated spatial correlation feature length. The longitudinally discrete monitoring sections of the guide embankment are reorganized into generalized evaluation units with dynamic spatial boundaries. And perform grid fusion processing on multi-source heterogeneous data within the cell, specifically including the following steps: Step S331: Determine the spatial boundary of the generalized evaluation unit. Using the longitudinal axis of the guide dike as... The axis is selected as the center coordinate, with the monitoring section where the most significant correlation drift or the largest structural response is currently occurring. Based on the spatial correlation feature length determined in step S323 Define a generalized evaluation unit Physical coverage : This physical range constitutes the dynamic evaluation range for structural toughness assessment at the current moment, and its scale adjusts as the system coupling strength changes. As the scale increases, the evaluation scope expands to both sides, incorporating more structural segments into the overall assessment.

[0067] Step S332: Identify valid monitoring nodes within the evaluation unit. Traverse the entire monitoring node set along the guide embankment to identify the set of sensor node indices falling within the aforementioned physical coverage area. : ;in, For the first The longitudinal coordinates of each monitoring section. This sensor node index set. A subset of data sources was determined to participate in the calculation of the unit attributes at the current moment.

[0068] Step S333: Perform data grid fusion and feature reconstruction. This is for the sensor node index set. The system incorporates multi-source monitoring data and employs a strategy combining the most unfavorable response envelope with environmental homogenization to fuse discrete measurement point data into a comprehensive evaluation unit. eigenvectors of the state For the significant wave height, considering that... Since sea state differences are relatively small within the scale range, the spatial arithmetic mean is used as the unit characteristic value: ,in, For sensor node index set The number of nodes in Vertical axis The measured effective wave height at the location. For the maximum scour depth at the toe of the dike and the cumulative horizontal displacement of characteristic points on the dike body, considering that structural failure is usually controlled by local extrema, a maximum value extraction strategy is adopted to reflect the most unfavorable safety state within the unit: ; ; The aforementioned grid fusion algorithm transforms point-based monitoring data, originally based on a single cross-section, into generalized evaluation units capable of representing a certain spatial scale. The overall state data. This feature vector It is then input into the toughness assessment module to calculate the failure probability of the structural system that takes into account the effects of spatial correlation drift.

[0069] This embodiment introduces the cumulative dissipation energy index. As the third-dimensional state variable describing the historical damage state of the guide embankment structure, it characterizes the irreversible energy consumed by the structure under long-term cyclic wave loads due to plastic deformation and frictional slip. The specific quantification process includes the following steps: Step S411: Calculate the equivalent wave impact force. Based on fluid dynamics principles, the time-varying effective wave height is calculated. This is converted into an equivalent horizontal wave force acting on the wave-facing surface of the guide breakwater structure. Considering the real-time nature of on-site monitoring, a simplified physical model is used for calculation: ;in, This is the wave force coefficient, and its value depends on the type of guide seawall revetment. The density of seawater; It is the acceleration due to gravity; The force characteristic width is taken here as the generalized evaluation unit determined in step S433. The unit width. This formula calculates... It represents the driving potential of the current environmental load on the work done by the structure.

[0070] Step S412: Calculate the incremental hysteretic dissipation energy. The structure undergoes displacement under wave forces, and its energy dissipation originates from friction between the foundation and the levee body, as well as plastic compression of internal pores. The horizontal cumulative displacement at key levee points obtained through monitoring is used to calculate this. Calculate the time step Displacement increment within Current moment Incremental hysteresis dissipation energy for: ;in, This is the hysteresis efficiency coefficient, used to correct the difference between theoretical work done and actual plastic energy dissipation, reflecting the stiffness degradation characteristics of the structure under cyclic loading. If If the energy dissipation increment at that moment approaches zero.

[0071] Step S413: Generate the normalized cumulative dissipation energy index. For the initial time... Up to the current moment All incremental hysteresis dissipation energy is accumulated, and the energy dissipation capacity of the structure is determined. After normalization, the dimensionless cumulative dissipation energy index is obtained. : Substituting and expanding, we get: ;in, The total energy absorbed by the structure from its initial state to complete loss of load-bearing capacity is represented by this parameter, obtained through push-down analysis of typical sections of the guide embankment or centrifuge model tests. The calculated cumulative energy dissipation index... The value range is [0,1], and this index is related to the maximum scour depth at the toe of the dike. Horizontal cumulative displacement of characteristic points of the embankment Together, they constitute a three-dimensional feature space describing the health status of the guide embankment.

[0072] This embodiment combines a physics-driven resilience attenuation model with a data-driven time-varying correlation structure to construct a model based on effective wave height. Maximum scouring depth at the toe of the dike and cumulative dissipation energy index The construction process of the three-dimensional dynamic vulnerability surface with independent variables includes the following steps: Step S421: Establish a non-stationary resistance attenuation model. The ultimate wave resistance of the guide dike structure decreases with structural physical damage and cumulative fatigue. In this embodiment, a time-varying median structural resistance is defined. The degradation law is described using a double exponential decay function: ;in, This represents the median resistance of the guide embankment structure in its initial intact state; The design limits the maximum allowable brush depth; For sea brush sensitivity coefficient, These two coefficients, representing energy dissipation sensitivity coefficients, were determined through regression analysis of historical failure case data or finite element numerical simulation inversion. The model quantifies the impact of structural physical damage on resistance to force decay.

[0073] Step S422: Calculate the time-varying comprehensive logarithmic standard deviation. The shape of the fragility curve depends on the uncertainty of the load and resistance and their correlation. Utilize the estimated values ​​of the time-varying parameter vector describing the correlation structure between variables, updated in step S234. Extract the correlation coefficient between load and structural response. (This coefficient is derived from the correlation parameter matrix) elements in and (Determined comprehensively). Construct the time-varying comprehensive logarithm standard deviation. : ;in, The logarithmic standard deviation of the structural resistance. This represents the logarithmic standard deviation of the wave load. The correlation coefficient introduced here... This reflects the dynamic impact of the degree of coupling between load uncertainty and resistance uncertainty on the overall reliability of the system under different sea states and damage stages.

[0074] Step S423: Generate the analytical expression for the three-dimensional dynamic vulnerability function. Assuming the structure-function follows a log-normal distribution, define the three-dimensional dynamic vulnerability function. The function outputs the value at a given effective wave height. Maximum scouring depth at the toe of the dike and cumulative dissipation energy index Probability of conditional failure under combination : ,in, Let be the standard normal cumulative distribution function. Substituting the decay model from step S421 into the above equation and expanding it using the properties of logarithmic operations, we obtain the analytical expression: From this formula, we can see the probability of condition failure. It is positively correlated with significant wave height, maximum scouring depth at the toe of the dike, and cumulative dissipated energy index, and its value is affected by the time-varying comprehensive logarithmic standard deviation. Adjustments.

[0075] Step S424: Discretize and construct the fragile surface data volume. Within the variable domain... Perform mesh generation and calculate the value at each grid point. The value is used to generate a three-dimensional discretized data volume. When the monitoring system inputs a real-time generalized evaluation unit... eigenvectors At that time, the current failure probability is obtained directly from the data volume using a trilinear interpolation algorithm.

[0076] This embodiment utilizes the constructed three-dimensional dynamic vulnerability surface, combined with the predicted distribution of future environmental loads, to establish a Markov process model, in order to quantitatively predict the evolution trend of the safety state of the guide embankment structure within a future time window. Specifically, it includes the following steps: Step S431: Construct the probability density function of future environmental loads. Based on meteorological and hydrological forecast data, obtain the prediction time window. Statistical characteristics of effective wave height within the range. Assuming the future wave height follows a Weibull distribution, a probability density function for predicting the wave height is constructed. : ,in, For wave height variable, These represent the shape, scale, and location parameters within the corresponding prediction window, respectively. These parameters are determined by the statistical moments output by the short-term wave forecast model using the method of moments estimation. This function defines the probability distribution characteristics of the environmental load intensity at future times.

[0077] Step S432: Calculate the state transition probability. Define the discrete safety state space of the guide embankment structure. ,in Indicates "fully functional state". This indicates a "failure state." The three-dimensional dynamic vulnerability function generated in step S423 is used. The computational structure is now in its intact state. Transfer to failure state single-step transition probability According to the law of total probability, this probability is the expected integral of the fragility function with respect to the load distribution: ,in, and This represents the maximum scour depth and cumulative energy dissipation index at the current monitoring moment. This integral calculation comprehensively considers the structure's current physical damage state, energy dissipation history, and the randomness of future environmental loads, outputting the probability of structural failure within a future window. Correspondingly, the probability of the structure remaining intact is... .

[0078] Step S433: Assemble the state transition probability matrix. This is based on the assumption of the irreversibility of the failure state, meaning that once the structure enters the failure state... It will not automatically return to its perfect state without manual repair. . Construction Time to State transition probability matrix : ; This matrix describes the dynamic evolution rules of the system within the prediction step.

[0079] Step S434: Predict the future state distribution vector. Define the state distribution vector. ,in and Representing time respectively The probability of the structure being in an intact or failed state. Based on the Markov chain property, predict future moments. State distribution vector : If at the current moment If the structure is confirmed to be in good condition, then the initial vector is taken as follows. Through this matrix operation, the system outputs the predicted failure probability value at the predicted time. This predicted value comprehensively considers the deformation of vulnerable surfaces caused by the drift of correlation parameters (through...). Time-varying composite log-standard deviation This approach, considering the uncertainties of environmental loads, enables a quantitative assessment of the safety risks of the guide embankment structure. For the aforementioned integral calculations, this embodiment employs the Gauss-Legendary quadrature formula or the trapezoidal numerical integration method for solution.

[0080] This embodiment constructs the system functional potential energy based on the probability index and physical state index obtained from the aforementioned calculations. This index maps the statistical reliability and physical remaining lifetime of the guide embankment structure to a unified energy scalar, serving as the basis for subsequent toughness analysis based on potential energy field theory. The specific calculation process includes the following steps: Step S511: Extract key system state parameters. Extract the current time step from the calculation results of the previous steps. The two core state variables: the condition failure probability output by step S423. This value reflects the instantaneous instability risk of the system under the current load and resistance coupling action; the cumulative dissipated energy index output by step S413 This value reflects the degree of irreversible physical damage accumulated throughout the system's history.

[0081] Step S512, construct the system's functional potential energy function. Define the system's functional potential energy. This refers to the ability of a guide embankment structure to resist external disturbances and maintain its designed function. Considering that the functional potential energy should increase with decreasing failure probability and reduced accumulated dissipated energy, this embodiment employs a product-type coupled model: ,in, The nominal functional potential energy upper limit (usually normalized to 1.0) represents the potential energy level of the structure in an ideal, intact, and risk-free state. It characterizes the reliability margin of the system; It characterizes the proportion of the system's remaining physical lifetime; and These are the reliability weight index and the damage weight index, respectively. This is used to adjust the relative importance of statistical risk and physical damage in potential energy assessment. For guide embankment structures where sudden brittle failure is of particular concern, the value can be increased. Value; for structures where long-term fatigue durability is a concern, the value should be increased. value.

[0082] Step S513: Determine the potential energy boundary conditions. To ensure the numerical stability of subsequent potential energy field calculations, the calculated boundary conditions are... Perform boundary constraint processing. When Approaching 1 (i.e., it will definitely fail) or When it approaches 1 (i.e., the energy consumption capacity is exhausted), let This indicates that the system's function has been completely lost, and its potential energy has dropped to its lowest potential energy state. The system's functional potential energy... The physical meaning of this value is that the higher the value, the greater the potential for the structure to retain functions to resist future disasters, and the higher the system's resilience redundancy.

[0083] This embodiment constructs a generalized repair potential flow. This is used as a rate index to describe the recovery capacity of an external intervention system injected into a guide embankment structure. This index quantifies the dynamic process by which repair resources are transformed into system functional recovery capacity under environmental constraints. The specific calculation includes the following steps: Step S521: Determine the baseline repair power. Based on the emergency response plan and available construction resources for the area where the guide dike project is located, quantify the theoretical maximum repair rate and define it as the baseline repair power. This parameter characterizes the increase in system functional potential energy per unit time under ideal sea conditions and abundant resources. The maximum value is determined by the product of the operating efficiency of a single construction shift and the number of devices that can operate simultaneously. For a specific type of guide dike structure, this value is considered the inherent upper limit of the engineering support system's capacity.

[0084] Step S522: Construct an environmental accessibility attenuation model. Actual repair operations are strictly constrained by sea state conditions. Utilize real-time monitored effective wave height... Construct environmental accessibility coefficient This describes the nonlinear reduction in efficiency caused by severe sea conditions at sea, expressed using a logistic function: ,in, The critical wave height for construction vessel operations; Sea state sensitivity coefficient This model is used to regulate the rate at which operational efficiency decreases with increasing wave height. It ensures that the repair potential flow approaches zero when the effective wave height exceeds the operational limit.

[0085] Step S523: Construct a resource allocation time lag model. Considering the emergency response activation time and material transportation cycle, define a resource allocation efficiency coefficient. This reflects the gradual availability of restoration resources over time, using a saturated growth model. ,in, The repair mechanism is triggered at the time when the system's functional potential energy is at its peak. (Times when the threshold is lower than the preset warning threshold) This is the time constant characteristic of resource allocation, representing the average time required for the first core rescue forces to arrive on-site and commence operations. hour, .

[0086] Step S524: Synthesize the generalized repair potential flow. Taking into account the above physical constraints, calculate the generalized repair potential flow injected into the system at the current moment. : Substituting and expanding, we get the complete expression: Generalized repair potential energy flow The physical meaning is: at the current moment After considering sea state obstruction and logistical delays, the actual energy injection rate that the external rescue system can provide to the breakwater structure is considered. In the resilience assessment equation, this term exists as a source term, interacting with the dissipation term that leads to potential energy decay, and together determining the evolution trajectory of the system's resilience.

[0087] Based on the aforementioned system functional potential energy and repair potential energy flow, this embodiment quantifies the comprehensive resilience index of the levee structure throughout the entire disaster-recovery process by constructing the state evolution equation and energy integral of the time-varying system. The specific calculation process includes the following steps: Step S531: Construct the system state dynamics evolution equation. After introducing an external repair mechanism, the system's damage state is jointly determined by the cumulative damage caused by environmental loads and the energy injection from repair resources. Define the net cumulative dissipation energy index after repair correction. This indicator reflects the time at time The actual degree of physical damage remaining in the structure: ,in, The cumulative dissipated energy index calculated in step S413 considering only the load effect; The maximum energy consumption capacity defined in step S413; This refers to the generalized repair potential flow calculated in step S524; This represents the triggering moment of the repair mechanism. The equation establishes a mathematical mapping relationship between the repair energy flow and the reversal of the structural physical damage state. Subsequently, the calculated... Replace the formula in step S512 The real-time system functional potential energy, taking into account the repair effect, was recalculated. .

[0088] Step S532: Perform Hamiltonian action integration. Integrate the system's functional potential energy. As an energy function describing the system state, within a preset evaluation time window By performing a definite integral within the inner, the toughness action quantity is obtained. : This represents the total functional capacity actually retained by the dike structure throughout the entire disaster and recovery cycle. The trapezoidal rule or Simpson's integral method is used for numerical calculation.

[0089] Step S533: Generate the comprehensive toughness index. To eliminate the influence of assessment time and structural scale, the toughness application rate is adjusted. After normalization, a dimensionless comprehensive resilience index is obtained. : ,in, The upper limit of the nominal functional potential energy defined in step S512; The total duration of the assessment period is used to calculate the comprehensive resilience index. The value range is [0,1]. This index comprehensively reflects three key characteristics of the dike system in responding to extreme disasters: resistance (by...) The rate of decline is reflected in the absorption capacity (by the rate of descent). The lowest trough value reflects) and resilience (by The upward slope is reflected in the trend. When the value is below the preset safety threshold, the system determines that the current structure's resilience redundancy or the preset repair resources are insufficient to cope with the predicted disaster risks.

[0090] This embodiment supports the operation of the above method through physical hardware deployment, specifically including: Deploy an edge data acquisition and preprocessing unit. A data acquisition controller (DAU) is deployed in the field, connecting to an acoustic Doppler wave meter, a GNSS receiver, and a multibeam scanning sonar. The DAU utilizes its built-in FPGA chip to perform hardware-level filtering and noise reduction on the raw signal and performs time stamp alignment based on the NTP protocol to ensure effective wave height. Horizontal cumulative displacement of the embankment and the maximum scouring depth at the toe of the dike Synchronization in the time dimension.

[0091] Construct a dual-redundant data transmission link. Establish a dual-channel link between the field DAU and the computing center: the main channel uses a fiber optic network to transmit high-bandwidth point cloud and wave data, while the backup channel uses 5G or BeiDou short message service. The system is configured with a breakpoint resume function; when communication is interrupted, data is temporarily stored locally on the DAU and retransmitted after recovery to ensure the continuity of historical samples required to construct the Copula function.

[0092] Configure a high-performance computing server cluster. Remote servers serve as the core algorithm carriers, including: CPU unit: responsible for logic control, task scheduling, and state transition probability matrix. Assembly and update; GPU acceleration unit: Configured with a GPU cluster supporting the CUDA architecture, utilizing parallel floating-point operations to accelerate Bayesian inference of time-varying correlation parameters, interpolation calculation of three-dimensional dynamic vulnerable surfaces, and resilience action. The numerical integral meets the timeliness requirements of engineering early warning; Storage unit: RAID disk array is used to store monitoring time-series data, Copula model parameters, and overall resilience index. Record.

[0093] Deploy a visualization terminal. The terminal connects to the server via a drive circuit to render the 3D model of the guide embankment in real time and visually display the system's functional potential. Generalized repair potential flow and comprehensive resilience index Evolutionary chart.

[0094] This embodiment provides a computer-readable storage medium, wherein the stored instructions, when executed by a processor, implement the aforementioned method, specifically controlling the computing device to perform the following operations: Data processing and parameter inference. Reading valid wave height. Horizontal cumulative displacement of the embankment and the maximum scouring depth at the toe of the dike Original data; a sliding window algorithm is used to extract samples, and Bayesian inference is used to iteratively solve for the time-varying correlation parameter vector. And store by timestamp.

[0095] Vulnerability analysis and discretization. Based on the median structural resistance. and time-varying combined logarithm standard deviation Constructing a three-dimensional dynamic vulnerability function Discretized data volumes, generated by dividing the data into three-dimensional meshes and stored in lookup tables, are used for quickly querying the probability of condition failure. .

[0096] Toughness calculation and integration. Based on baseline repair power. Calculating the generalized repair potential flow Update system functional potential The state value is then used to perform numerical integration on the potential energy curve within the specified window to obtain the toughening effect. and comprehensive resilience index .

[0097] Dedicated data structure for storage. Specific regions are allocated for storing computational data: the state transition probability matrix is ​​stored in a sparse matrix format. To optimize memory usage; to establish support for random access based on time indexes. Historical time-series files to support trend retrospective analysis.

Claims

1. A method for evaluating the toughness of a breakwater structure and predicting the probability of damage state, characterized by, Includes the following steps: S1. Collect multi-source heterogeneous monitoring data, preprocess the collected data, and construct a standardized state monitoring vector sequence; S2. Based on the standardized state monitoring vector sequence, construct a damage mode coupling model based on the time-varying Copula function, and output the correlation coefficient matrix that evolves over time; S3. Obtain the maximum correlation coefficient modulus in the correlation coefficient matrix, calculate the length of the damage space related features based on the change of the maximum correlation coefficient modulus, and generate a generalized evaluation unit based on the length of the damage space related features. S4. Within the generalized evaluation unit, a three-dimensional dynamic vulnerability model considering accumulated energy dissipation is constructed, and the instantaneous probability distribution vector of the guide embankment structure at each damage level is predicted based on the three-dimensional dynamic vulnerability model. S5. Calculate the system functional potential energy based on the instantaneous probability distribution vector, and map the emergency repair resources into a generalized repair potential energy flow. By performing a weighted integral on the system functional potential energy and the generalized repair potential energy flow, output a comprehensive resilience index for quantitative evaluation criteria and decision-making basis.

2. The method for evaluating the toughness of a guide embankment structure and predicting the probability of damage states according to claim 1, characterized in that, The vector elements in the standardized state monitoring vector sequence include significant wave height, real-time tide level, maximum scour depth at the toe of the dike, cumulative horizontal displacement of characteristic points on the dike body, and pore water pressure. The preprocessing specifically includes: using the Kalman filter algorithm to denoise the original time series, and using linear interpolation to align data with different sampling frequencies to a uniform time step.

3. The method for evaluating the flexibility of a breakwater structure and predicting the damage state probability according to claim 1, characterized in that, The S2 step specifically includes: The failure modes of the guide dike structure are identified based on the standardized state monitoring vector sequence, including the instability of the revetment block, toe scouring and dike slippage, and the edge cumulative distribution of each failure mode is fitted. A damage mode coupling model based on the time-varying Copula function is used to describe the joint probability distribution among the failure modes; Historical monitoring data is extracted using a sliding time window technique, and the posterior distribution of the time-varying parameter vector describing the correlation structure between variables is updated using a Bayesian inference algorithm. Then, the correlation coefficient matrix between each failure mode at the current time is calculated.

4. The method for evaluating the flexibility of a breakwater structure and predicting the damage state probability according to claim 1, characterized in that, The step of calculating the length of the relevant feature in the damage space based on the change in the correlation coefficient specifically includes: Obtain the maximum correlation coefficient magnitude in the correlation coefficient matrix, whereby the maximum correlation coefficient magnitude characterizes the coupling strength between different failure modes at the current moment; The length of the relevant features in the damage space is calculated using a preset nonlinear mapping function; The nonlinear mapping function is configured such that the length of the damage space-related feature increases nonlinearly between the preset base length and the maximum allowable evaluation length as the maximum correlation coefficient modulus increases.

5. The method for evaluating the toughness of a guide embankment structure and predicting the probability of damage state according to claim 4, characterized in that, The steps for generating a generalized evaluation unit based on the length of the damage space-related features specifically include: Based on the longitudinal axis of the guide embankment, determine whether the geometric center distance between adjacent initial discrete micro-elements is less than the length of the damage space related feature; The adjacent initial discrete micro-elements whose geometric center distance meets the condition are meshed to generate the generalized evaluation unit, and the equivalent physical parameters of the generalized evaluation unit are redefined as the physical evaluation boundary for vulnerability analysis.

6. The method for evaluating the toughness of a guide embankment structure and predicting the probability of damage state according to claim 1, characterized in that, Before constructing the three-dimensional dynamic vulnerability model, the following steps are also included: Calculate the historical cumulative dissipation energy index of the guide embankment structure, which is defined as the ratio of the cumulative plastic deformation energy of the structure in the hysteresis cycle to the design ultimate energy storage capacity.

7. The method for evaluating the toughness of a guide embankment structure and predicting the probability of damage state according to claim 6, characterized in that, The steps for constructing a three-dimensional dynamic vulnerability model that considers accumulated energy dissipation specifically include: The predicted load intensity at the next moment and the historical cumulative dissipation energy index are introduced as independent variables into the vulnerability function; Within the generalized evaluation unit, a three-dimensional dynamic vulnerability model is constructed that includes the predicted load intensity, the historical cumulative dissipation energy index, and the damage state. The vulnerability function includes a median resistance that decreases exponentially with the cumulative dissipation energy, and the median resistance decreases power-lawly with the increase of the historical cumulative dissipation energy index.

8. The method for evaluating the toughness of a breakwater structure and predicting the damage state probability according to claim 7, characterized in that, The steps for predicting the instantaneous probability distribution vector of the guide embankment structure at each damage level specifically include: A non-homogeneous state transition matrix is ​​generated based on the aforementioned three-dimensional dynamic vulnerability model; The instantaneous probability distribution vectors of the guide embankment structure at each of the following time steps—intact, slightly damaged, moderately damaged, and severely damaged—are calculated using the non-homogeneous state transition matrix.

9. The method for evaluating the toughness of a guide embankment structure and predicting the probability of damage state according to claim 1, characterized in that, The system functional potential energy is a weighted sum of the instantaneous probability distribution vectors of each damage and the corresponding functional retention coefficients of the generalized repair potential energy flow; The step of mapping emergency repair resources to a generalized repair potential energy flow specifically includes: Based on preset resource conversion efficiency, rate saturation factor, and damage state threshold for initiating repair, the emergency repair resource reserve and mobilization rate are converted into the generalized repair potential energy flow that performs work on the damaged system.

10. The method for evaluating the toughness of a guide embankment structure and predicting the probability of damage state according to claim 9, characterized in that, The specific steps for outputting the comprehensive resilience index include: Within a predetermined evaluation period, the product of the system functional potential energy and the functional maintenance weight coefficient, and the product of the generalized repair potential energy flow and the repair resource weight coefficient are weighted and summed. The result of the weighted sum is then integraled over the time domain of the evaluation period to obtain the comprehensive resilience index.

Images