A method and system for beach restoration decision-making based on multi-source data

Abstract

The present application relates to the technical field of coastal engineering, and discloses a kind of beach restoration decision-making method and system based on multi-source data, its method includes the following steps, obtains the multi-source data of beach monitoring, carries out data fusion processing, generates the feature tensor of unified space-time resolution;Symbol regression analysis is carried out based on feature tensor, and the control equation of beach evolution is automatically found;Evolution equation is input into decision optimization framework, and beach restoration decision scheme is generated.The present application realizes the accurate characterization of beach evolution process and the intelligent generation of restoration decision by the organic combination of multi-source data fusion, symbol regression analysis and decision optimization framework;The scheme can make full use of the information of various monitoring data, automatically find the control equation of beach evolution and its time-varying characteristics, and directly convert physical law into specific engineering decision suggestion, so as to improve the prediction accuracy and decision efficiency of beach restoration.

Description

Technical Field

[0001] This invention relates to the field of coastal engineering, and more specifically, to a method and system for beach restoration decision-making based on multi-source data. Background Technology

[0002] In the field of coastal engineering, shoreline restoration is a crucial measure for maintaining coastline stability and ecological functions. Traditional shoreline restoration decision-making primarily relies on expert experience and numerical simulation models, such as the STWAVE model for wave field simulation, the GENESIS model for shoreline evolution prediction, and the XBeach model for profile response analysis. While these models are widely used in engineering practice, they still face the following technical challenges:

[0003] First, shoreline evolution involves multi-source heterogeneous data such as remote sensing images, acoustic bathymetry data, GPS shoreline measurements, wave buoy monitoring, and tide gauge records. These data have significant differences in spatiotemporal resolution, data format, and physical meaning. Traditional numerical models cannot directly integrate and utilize all data sources and can usually only selectively use some data, resulting in insufficient information utilization.

[0004] Secondly, traditional numerical models such as STWAVE and GENESIS are based on pre-defined physical equations, which are mostly derived from idealized physical assumptions and ignore complex physical processes such as the nonlinear coupling between storm surges and waves, and the multi-scale effects of sediment transport. For example, in the Shidao Bay beach restoration project, traditional models struggled to accurately capture the complex wave shadow zone effect generated by the interaction between southeast-facing waves and the breakwater.

[0005] Furthermore, the parameters and structure of the governing equations of the beach system will change over time with seasonal variations and extreme weather events. Traditional static models require manual adjustment of parameters to adapt to environmental conditions at different times, lacking adaptive capabilities and affecting the accuracy of predicting long-term restoration effects.

[0006] Finally, even if wave field distribution and shoreline evolution predictions are obtained through traditional models, it is still necessary to compare a large number of schemes (such as the 14 schemes in the Shidaowan project) to determine the optimal repair measures. There is a lack of a systematic conversion method that directly transforms physical evolution laws into decision-making suggestions, resulting in low efficiency and reliance on experience-based judgment in the decision-making process. Summary of the Invention

[0007] This invention provides a beach restoration decision-making method and system based on multi-source data, which solves the technical problem that traditional beach restoration decision-making methods have limitations in multi-source data fusion and utilization, complex physical process simulation, time-varying characteristic adaptation and decision-making efficiency, and are difficult to meet the needs of engineering practice.

[0008] This invention provides a beach restoration decision-making method based on multi-source data, comprising the following steps:

[0009] S100 acquires multi-source data from beach monitoring, performs data fusion processing, and generates a feature tensor with unified spatiotemporal resolution.

[0010] The multi-source data includes remote sensing imagery data, acoustic bathymetry data, GPS shoreline measurement data, wave buoy monitoring data, and tide gauge station records.

[0011] S200, based on symbolic regression analysis using characteristic tensors, automatically discovers the governing equations of beach evolution;

[0012] S300 inputs the evolution equation into the decision optimization framework to generate beach restoration decision schemes;

[0013] The decision-making scheme for shoreline restoration includes the amount of artificial sand replenishment, the length of the sand-retaining dike, the elevation of the submerged dike, the number of groynes, and the inclination angle of the groynes.

[0014] Furthermore, S100 specifically includes the following steps:

[0015] S110, multi-source data acquisition and preprocessing, acquiring remote sensing image data, including visible light and infrared band images, performing atmospheric and geometric correction, acquiring acoustic bathymetry data, performing noise filtering and depth calibration, acquiring GPS shoreline measurement data, performing coordinate transformation and outlier removal, acquiring wave buoy monitoring data, including quality control of wave height, period and direction data, acquiring tide gauge station record data, performing tide datum unification and data supplementation;

[0016] S120, spatiotemporal alignment processing, uses bilinear interpolation to resample spatial data to a 5-meter grid resolution, uses cubic spline interpolation to resample time series to an hourly time scale, establishes a unified spatiotemporal reference system, ensures that all data sources are aligned in the same coordinate system, performs data synchronization checks, and ensures that timestamps from different data sources correspond accurately.

[0017] S130, physical characteristic calculation, uses the finite difference method to calculate the spatial gradient of topographic elevation, uses the central difference scheme to calculate the time derivative of wave height, calculates the divergence of sediment flux based on the finite volume method, and constructs composite physical quantities such as wave-tide coupling term and wave steepness-water depth ratio.

[0018] S140, Feature Tensor Generation: Design the data structure of the feature tensor, which includes spatial, temporal, and physical quantity dimensions. Fill the corresponding positions of the tensor with the aligned original observation data, integrate the calculated physical derivative features into the tensor, and perform data integrity checks and outlier handling.

[0019] Furthermore, S200 specifically includes the following steps:

[0020] S210, Data preprocessing and initialization, standardizes the feature tensor to unify the dimensions of each physical quantity, sets the candidate function set, initializes the genetic programming algorithm parameters, and constructs the fitness function;

[0021] S220, a genetic programming evolutionary process, uses a tree-structured encoding method to represent candidate equations and performs crossover and mutation operations;

[0022] S230, Sparse Regression Optimization, transforms the candidate equation terms generated by genetic programming into a coefficient matrix and applies the LASSO algorithm for sparsification.

[0023] S240, physical consistency verification, checking the dimensional consistency of the equations, verifying the law of conservation of sediment mass, and verifying the law of conservation of wave energy;

[0024] S250, time-varying feature analysis, uses a sliding window to segment the data sequence, identifies seasonal changes and extreme event periods, performs independent symbolic regression on each time period, generates a piecewise evolution equation, and establishes a smooth transition mechanism between time periods;

[0025] S260, Causal Structure Identification, uses the PC algorithm for causal discovery, calculates the structural edit distance of the DAG between adjacent periods, applies the PELT algorithm to detect structural abrupt points, constructs variable combination rules based on the DAG, and updates the candidate set of equation terms;

[0026] S270, Coupling Effect Analysis, calculates the cross-correlation function of variable pairs, generates first-order and second-order coupling terms, and constructs composite physical quantities.

[0027] Furthermore, the candidate function set includes: basic arithmetic operations; trigonometric functions; exponential functions; and power functions.

[0028] Furthermore, the parameters of the genetic programming algorithm include population size, number of iterations, crossover probability, and mutation probability.

[0029] Furthermore, in S220, the crossover operation includes randomly selecting two parent trees, randomly selecting the crossover point, and exchanging subtrees to generate a new individual.

[0030] Furthermore, in S220, the mutation operation involves randomly selecting mutation points and replacing them with new operators / variables with probability, using a tournament selection method for iterative evolution.

[0031] Furthermore, in S300, the following steps are specifically included: S310, Repairing engineering parameter settings: Determining the set of control variables, setting constraints, establishing parameter association constraints, and generating an initial solution set that satisfies the constraints;

[0032] S320, Evolutionary Prediction Simulation: Substitute the parameters of the repair scheme into the evolution equation;

[0033] S330, Construction of multi-objective optimization function: Combining the definition of beach stability index, engineering cost function and ecological impact index to construct a comprehensive evaluation function;

[0034] S340, Genetic Algorithm Optimization Solution: The repair scheme is represented by real number encoding, the algorithm parameters are set, simulated binary crossover and polynomial mutation are performed, and fast non-dominated sorting is adopted;

[0035] S350, Solution Evaluation and Screening: Analyze the characteristics of the optimal solution set, evaluate the feasibility of the project, calculate the cost-benefit ratio, and consider decision preferences;

[0036] S360, Uncertainty Analysis: Using the Monte Carlo method, random samples are generated, evolution results are calculated, statistical confidence intervals are obtained, and robustness indices of the scheme are calculated;

[0037] S370, Remediation Target Identification: Calculate the rate of shoreline change, identify erosion and sedimentation zones, assess topographic stability, and determine control sections;

[0038] S380, Optimization of Intervention Timing: Analyze seasonal changes, identify state transition points, design and implement solutions based on state prediction models, allocate construction timelines for phased projects, and optimize project progress;

[0039] S390, Decision Recommendation Generation: Determine project combinations, formulate layout plans, plan implementation sequence, compile construction guidelines, and output decision reports.

[0040] Furthermore, in S350, the analysis of the optimal solution set characteristics includes calculating the uniformity of solution distribution, evaluating the diversity index of solutions, and identifying key parameter combination patterns; evaluating engineering feasibility includes construction difficulty index, material availability, and environmental adaptability; calculating cost-benefit ratio includes stability improvement, engineering cost, and maximum allowable cost, and selecting the optimal solution.

[0041] This invention also proposes a beach restoration decision-making system based on multi-source data, comprising:

[0042] Data acquisition and preprocessing module: responsible for acquiring and preprocessing multi-source monitoring data, performing spatiotemporal alignment processing on the data, unifying it to a 5-meter grid resolution and an hourly time scale, calculating physical features and generating feature tensors containing multi-scale physical process information;

[0043] Evolutionary equation discovery module: Based on feature tensors, symbolic regression analysis is performed to automatically discover the governing equations of beach evolution through genetic programming algorithms, sparse regression optimization and physical consistency verification are performed, time-varying features and causal structures are analyzed, key coupling effects are identified, and finally evolutionary equations that can describe complex physical processes are generated.

[0044] Decision optimization module: Input the evolution equation into the decision optimization framework, set the restoration project parameters and perform evolution prediction simulation, construct a multi-objective optimization function, solve the optimal solution through a genetic algorithm, perform scheme evaluation, uncertainty analysis and restoration target identification, optimize the timing of intervention, and finally generate a complete beach restoration decision scheme.

[0045] The beneficial effects of this invention are as follows:

[0046] This invention achieves precise characterization of the shoreline evolution process and intelligent generation of restoration decisions through the organic combination of multi-source data fusion, symbolic regression analysis, and a decision optimization framework. This scheme fully utilizes information from various monitoring data to automatically discover the governing equations and time-varying characteristics of shoreline evolution, directly translating physical laws into specific engineering decision recommendations. This improves the prediction accuracy and decision-making efficiency of shoreline restoration, reducing reliance on expert experience. Furthermore, the scheme exhibits strong adaptability and robustness, capable of handling complex hydrodynamic conditions and extreme weather events, providing strong support for scientific decision-making in shoreline restoration projects. Attached Figure Description

[0047] Figure 1 This is a flowchart of a beach restoration decision-making method based on multi-source data proposed in this invention;

[0048] Figure 2 This is a structural block diagram of a beach restoration decision system based on multi-source data proposed in this invention.

[0049] In the diagram: 101, Data Acquisition and Preprocessing Module; 102, Evolution Equation Discovery Module; 103, Decision Optimization Module. Detailed Implementation

[0050] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0051] like Figure 1As shown, a beach restoration decision-making method based on multi-source data includes the following steps:

[0052] S100 acquires multi-source data from beach monitoring, performs data fusion processing, and generates a feature tensor with unified spatiotemporal resolution.

[0053] In one embodiment of the present invention, the following steps are specifically included:

[0054] S110, Multi-source Data Acquisition and Preprocessing: Acquire remote sensing image data, including visible and infrared band images, and perform atmospheric correction (using the FLAASH model) and geometric correction (RMS error <0.5 pixels); acquire acoustic bathymetry data, and perform noise filtering (using median filtering, 3×3 window size) and depth calibration (based on known control points, accuracy ±0.1m); acquire GPS shoreline measurement data, and perform coordinate transformation (using a 7-parameter transformation model) and outlier removal (3σ criterion); acquire wave buoy monitoring data, including quality control of wave height (Hs), period (T), and direction (θ) data (based on physical constraints); acquire tide gauge station record data, and perform tide datum unification (using the 85 elevation datum) and data completion (based on harmonic analysis method);

[0055] S120, spatiotemporal alignment processing, uses bilinear interpolation to resample spatial data, unifying it to a 5-meter grid resolution:

[0056] ;

[0057] ;

[0058] ;

[0059] in, This is the interpolation result at the target point (x, y). , , , Given the coordinates of four surrounding known points, For the interpolation coefficients in the x-direction, , For the interpolation coefficients in the y-direction, ;

[0060] Cubic spline interpolation was used to resample the time series data to a uniform hourly time scale.

[0061] ;

[0062] in, This is the interpolation result at time t. , , , The interpolation coefficients for the i-th segment are determined by the boundary conditions. For the i-th known time point, , For the time points to be interpolated, ;

[0063] Establish a unified spatiotemporal reference system to ensure that all data sources are aligned in the same coordinate system (using the CGCS2000 coordinate system), and perform data synchronization checks to ensure that the timestamps of different data sources correspond accurately (time error <1min).

[0064] S130, Physical feature calculation, using the finite difference method to calculate the spatial gradient of terrain elevation:

[0065] ;

[0066] ;

[0067] in, Let (i, j) be the terrain elevation at grid point (i, j). , Let x and y be the spatial step sizes. For grid coordinates, ;

[0068] The time derivative of the wave height was calculated using the central difference scheme:

[0069] ;

[0070] in, Let be the effective wave height at time t. For time step, For a point in time;

[0071] Calculation of divergence of sediment flux based on the finite volume method:

[0072] ;

[0073] ;

[0074] in, For sediment flux, , Let x and y be the components of the sediment flux. This represents the sediment transport coefficient. For the effective wave height, The angle of incidence of the wave. ;

[0075] Constructing composite physical quantities such as wave-tide coupling terms and wave steepness-depth ratio:

[0076] Coupling terms:

[0077] ;

[0078] in, The coupling coefficient is... For the effective wave height, The tide level;

[0079] Wave steepness ratio:

[0080] ;

[0081] in, For wave steepness ratio, For the effective wave height, For wavelength, For water depth;

[0082] S140, Feature Tensor Generation: Designing the Data Structure for Feature Tensors ,Include:

[0083] Spatial dimension ( ): 5m×5m grid, time dimension : 1-hour interval, physical quantity dimension : 12 variables;

[0084] Fill the corresponding positions in the tensor with the aligned original observation data:

[0085] ;

[0086] in, The feature tensor has dimensions [1000, 1000, 8760, 12]. For spatial grid indexing, , For time indexing, , For physical quantity indexing, , For the observation data mapping function, For spatial coordinates, For a point in time, Let l be the l-th physical quantity;

[0087] Integrate the calculated physical derivative features into the tensor:

[0088] ;

[0089] in The feature tensor has dimensions [1000, 1000, 8760, 12]. Let m be the index of the derivative feature, where m ∈ [1, 12]. ， For the index of the original features, ∈[1, 12], Spatial coordinates;

[0090] Perform data integrity checks (missing rate <5%) and outlier handling (based on the MAD method);

[0091] ;

[0092] in, For data points, the units are the same as the original variables. For the dataset, the units are consistent with the original variables. The median operation generates a final feature tensor (dimension: 1000×1000×8760×12) containing information about multi-scale physical processes.

[0093] S200, based on symbolic regression analysis using characteristic tensors, automatically discovers the governing equations of beach evolution;

[0094] In one embodiment of the present invention, the following steps are specifically included:

[0095] S210, Data preprocessing and initialization, standardizes the feature tensors to unify the dimensions of various physical quantities:

[0096] ;

[0097] in, For standardized data, The original data, The mean, Standard deviation;

[0098] Define the candidate function set ,include:

[0099] Basic arithmetic operations: ;

[0100] Trigonometric functions: ;

[0101] Exponential function: ;

[0102] Power function: , ;

[0103] Initialize genetic programming algorithm parameters: population size =500, number of iterations =1000, crossover probability =0.8, mutation probability =0.1;

[0104] Construct the fitness function:

[0105] ;

[0106] in =0.7, =0.2, =0.1 is the weighting coefficient. Mean square error, For the equation complexity, The degree of violation of physical constraints.

[0107] S220, the genetic programming evolutionary process, uses a tree-structured encoding method to represent candidate equations:

[0108] Internal nodes ∈ (Operator set)

[0109] Leaf node ∈ {variable set ∪ constant set}

[0110] Perform crossover operation:

[0111] Randomly select two parent trees , ;

[0112] Randomly select intersection points , ;

[0113] Swap subtrees to generate new individuals , ;

[0114] Perform mutation operation: randomly select mutation points. With probability Replace with new operators / variables;

[0115] The tournament selection method is used: k individuals (k=3) are randomly selected, and the individual with the highest fitness is selected to enter the next generation. The evolution is iterated until the termination condition is reached.

[0116] Termination condition: The number of iterations reaches T or the optimal fitness shows no improvement for 50 consecutive generations;

[0117] S230, sparse regression optimization, transforms candidate equation terms generated by genetic programming into a coefficient matrix. Sparsity reduction is performed using the LASSO algorithm:

[0118] ;

[0119] in, For the target variable, Let be the vector of coefficients to be determined. For regularization parameters ( =0.01), the complexity of the equation is controlled by the L1 regularization term:

[0120] ;

[0121] in For the i-th component of the coefficient vector, remove terms with small contributions: <0.001, thus obtaining the simplified equation expression;

[0122] S240, Physical consistency verification, checking the dimensional consistency of the equation: [left side] = [right side], verifying the law of conservation of sediment mass:

[0123] ;

[0124] in Because of the water depth, For sediment flux;

[0125] Verify the law of conservation of wave energy:

[0126] ;

[0127] in, Wave energy density, For group velocity, It is a dissipation term;

[0128] S250, time-varying feature analysis, uses a sliding window to segment the data sequence: window width w = 90 days, sliding step s = 30 days, to identify seasonal variations and extreme event periods; wavelet transform is used to analyze periodic features, and the POT method is used to identify extreme events. Sign regression is performed independently for each time period; the S220-240 method is applied to generate piecewise evolution equations.

[0129] ;

[0130] ;

[0131]

[0132] in, Location of the shoreline. For time, The evolution function for each time period is given by / s. For time intervals;

[0133] Establish a smooth transition mechanism between time periods: use a weighting function w(t) to achieve smooth transition;

[0134] S260, causal structure identification, uses the PC algorithm for causal discovery:

[0135] Significance level α = 0.05, maximum condition set size k = 3

[0136] Calculate the structural edit distance of a DAG between adjacent periods:

[0137] ;

[0138] in, , For edge set, This represents the set difference operation;

[0139] Applying the PELT algorithm to detect structural abrupt changes: penalty parameter β=0.1, minimum segment length l=30.

[0140] Based on DAG, construct variable combination rules: direct causal relationship → first-order term; indirect causal relationship → higher-order term; update the candidate set of equation terms.

[0141] S270, Coupling effect analysis, calculating the cross-correlation function of variable pairs:

[0142] ;

[0143] in, , It is a time series. , The mean, , Standard deviation For time delay, For the expected operation, The cross-correlation coefficient;

[0144] Generate first-order and second-order coupling terms:

[0145] First-order terms: ;

[0146] Second-order term: ;

[0147] Constructing composite physical quantities:

[0148] Wave height-tide level coupling term: ;

[0149] Wave steepness-depth ratio: ;

[0150] Assess the importance of each term: Through partial correlation analysis and variance contribution rate, generate the final composite equation, which is a complete evolution equation containing significant coupling terms;

[0151] S300 inputs the evolution equation into the decision optimization framework to generate beach restoration decision schemes;

[0152] In one embodiment of the present invention, the following steps are specifically included:

[0153] S310, Repair engineering parameter settings, determine the set of control variables;

[0154] ;

[0155] in, This refers to the amount of sand added artificially. ∈[1×10 4 5×10 5 ], The length of the sand-retaining dike. ∈[100, 500], The elevation of the submerged dike. ∈[1, 5], For the number of groynes. ∈[1, 5], The inclination angle of the groyne. ∈[30, 90];

[0156] Set constraints:

[0157] Total project cost ;

[0158] Ecological impact indicators ;

[0159] Spacing between structures ;

[0160] in The maximum permissible project cost, ecological impact indicators, among which To minimize the permissible ecological impact, the spacing between structures, among which Spacing between adjacent structures;

[0161] Establish parameter association constraints:

[0162] Ratio of dam length to sand replenishment volume ;

[0163] ratio of submerged dike elevation to water depth ;

[0164] Generate an initial solution set that satisfies the constraints. ,in The initial number of solutions;

[0165] S320, evolutionary prediction simulation, substitutes the repair scheme parameters into the evolutionary equation:

[0166] ;

[0167] in, Location of the shoreline. Because of the water depth, For the effective wave height, The wave direction angle, For tide level, For time;

[0168] Set the simulation time step =1 hour, total duration =10 years, using the fourth-order Runge-Kutta method to solve the evolution equations:

[0169] ;

[0170] ;

[0171] ;

[0172] ;

[0173] ;

[0174] Among them, key moments are recorded. , , ..., shoreline location ;

[0175] Calculate evolutionary characteristic indicators:

[0176] shoreline change rate ;

[0177] Stability Index ;

[0178] Wave attenuation coefficient ;

[0179] in, This represents the change in shoreline. The maximum permissible variation is the wave attenuation coefficient, where... , These are the incident and transmitted wave heights, respectively.

[0180] S330, multi-objective optimization function construction, defining the beach stability index:

[0181] ;

[0182] in, This represents the change in shoreline. This is the maximum permissible variation.

[0183] Define the project cost function:

[0184] ;

[0185] in, Cost per unit of sand replenishment The unit price of the sand-retaining dike. The length of the sand-retaining dike. The unit price of the submerged dike. The length of the submerged dike; ;

[0186] Define ecological impact indicators:

[0187] ;

[0188] in, These are the weighting coefficients. These are the sub-items of ecological impact factors;

[0189] Set the weight coefficients for the objective function: =0.4 (stability) =0.3 (cost) =0.3 (ecology);

[0190] Construct a comprehensive evaluation function:

[0191] ;

[0192] S340, Genetic Algorithm Optimization Solution: The repair scheme is represented using real-number encoding.

[0193] ;

[0194] in, As a repair solution, This refers to the amount of sand added artificially. The length of the sand-retaining dike. The elevation of the submerged dike. For the number of groynes. The inclination angle of the groyne;

[0195] Set algorithm parameters: population size =200, number of iterations =500, crossover probability =0.8, mutation probability =0.1;

[0196] Perform a simulated binary cross (SBX):

[0197] ;

[0198] ;

[0199] Where, β= , ∈[0,1], =20, distribution index , For the parent generation, , For offspring individuals;

[0200] Perform polynomial mutation:

[0201] ;

[0202] in, , ∈[0,1], =20, distribution index For individuals to be mutated, , The upper and lower boundaries;

[0203] Using fast non-dominated sort: Calculate the dominance number and the dominated solution set according to Layer, update crowding distance, update Pareto optimal solution set. ;

[0204] S350, Solution Evaluation and Screening:

[0205] Analyze the characteristics of the optimal solution set: calculate the uniformity of solution distribution. Evaluation of solution diversity indicators Identify key parameter combination patterns;

[0206] Assessing project feasibility: Construction difficulty index (CI), material availability index (MI), and environmental adaptability index (EI);

[0207] Calculate the cost-benefit ratio:

[0208] ;

[0209] in, To improve stability, For project costs, The maximum allowable cost;

[0210] Consider decision preferences:

[0211] Utility function:

[0212] ;

[0213] in, These are the weighting coefficients. , For utility function, For the decision variable, with the same units as the original variable, select the optimal solution. ;

[0214] S360, Uncertainty Analysis, using the Monte Carlo method: Generate =1000 sets of parameter samples, parameter perturbation range ±20%;

[0215] Generate random samples: ;

[0216] in, The optimal solution is... For random perturbation, It is the square of the variance;

[0217] Calculation results of evolution: Statistical confidence interval: [ (t), [t], calculate the robustness index of the scheme: ;

[0218] S370, target point identification for repair, calculation of shoreline change rate: Identify erosion and sedimentation zones, especially those with strong erosion: <-2m / year; strong sedimentation: >1m / year.

[0219] Assess terrain stability:

[0220] ;

[0221] in, For terrain elevation, The Laplace operator for elevation, It is the terrain stability index;

[0222] Determine the control section: the location of the maximum erosion rate, the point of topographic abrupt change, and the area of ​​concentrated hydrodynamic forces.

[0223] S380, Optimization of Intervention Timing, Analysis of Seasonal Variations: Wave High Cycle tide period sediment transport cycle Identify state transition points: eigenvalues The zero point, the inflection point of the phase space trajectory;

[0224] Predicting the critical moment: ;

[0225] in, At the initial moment, For predicting time intervals, This is the critical moment;

[0226] Based on the state prediction model, design and implement the project, allocate construction time in phases, and optimize the project schedule to minimize the construction period. Maximize work efficiency η;

[0227] S390, Decision suggestion generation, determining project combination: main project type, auxiliary measures configuration, monitoring system deployment, formulating layout plan: site layout drawing, section design drawing, construction zoning drawing, planning implementation sequence: total construction period. Sub-item construction period Construction sequence, preparation of construction guidelines: technical standards, quality requirements, safety measures, output of decision reports: project implementation plan, expected effects, and risk assessment.

[0228] like Figure 2 As shown, a beach restoration decision-making system based on multi-source data includes the following modules:

[0229] Data Acquisition and Preprocessing Module 101: Responsible for acquiring and preprocessing multi-source monitoring data, including remote sensing imagery, acoustic bathymetry, GPS shoreline measurements, wave buoy data, and tide gauge data. It performs spatiotemporal alignment processing on the data, unifying it to a 5-meter grid resolution and an hourly time scale. It calculates physical characteristics and generates feature tensors containing multi-scale physical process information.

[0230] Evolutionary Equation Discovery Module 102: Based on feature tensors, it performs symbolic regression analysis and automatically discovers the governing equations of beach evolution using a genetic programming algorithm. It then performs sparse regression optimization and physical consistency verification, analyzes time-varying characteristics and causal structures, identifies key coupling effects, and ultimately generates evolutionary equations that describe complex physical processes.

[0231] Decision Optimization Module 103: Inputs the evolutionary equations into the decision optimization framework, sets the restoration engineering parameters, and performs evolutionary prediction simulations. It constructs a multi-objective optimization function and solves for the optimal solution using a genetic algorithm. It then performs scheme evaluation, uncertainty analysis, and restoration target identification, optimizes the timing of intervention, and ultimately generates a complete beach restoration decision scheme.

[0232] An example of steps S100-S300 of the beach restoration decision-making method based on multi-source data is given in one embodiment of the present invention, using the steps of the above-described beach restoration decision-making method based on multi-source data and the modules in the system:

[0233] S100, Multi-source Data Acquisition and Preprocessing: In the Shidaowan Coastal Zone Protection and Restoration Project, the specific implementation steps for multi-source data acquisition and preprocessing are as follows:

[0234] Marine dynamic environment data acquisition: Collect marine dynamic environment observation data such as tide level, wave, and ocean current in Shidao Bay. Use a combination of automatic observation stations and field surveys to obtain continuous marine dynamic environment parameters, establish a tidal model for Shidao Bay, and obtain the tidal level change pattern.

[0235] Topographic and geomorphological data acquisition: Underwater topographic data of Shidao Bay are acquired through a multibeam echo sounder system, and the beach topography is measured using RTK-GPS technology to obtain high-precision shoreline location and beach elevation data. Historical topographic data, including historical nautical charts and measurement data, are also collected.

[0236] Remote sensing image data acquisition: acquire multi-temporal high-resolution satellite images (such as GF-1, GF-2, etc.) of the Shidao Bay area, collect drone aerial images, obtain detailed features of the shoreline, and establish a shoreline change sequence using historical remote sensing image data;

[0237] Sediment transport data collection: Sediment sampling points are set up at key sections to obtain parameters such as sediment particle size and content, measure suspended sediment concentration and transport flux, and analyze sediment sources and transport paths;

[0238] Socioeconomic data collection: Collect socioeconomic data such as population, industrial structure, and land use in the area surrounding Shidao Bay; investigate the current status and planning of coastal zone development and utilization; and assess the value of coastal zone resources and development intensity.

[0239] The collected data are shown in Table 1.

[0240] Table 1. Statistical Table of Multi-Source Data Collection on Shidao Bay Coastal Zone

[0241]

[0242] Data preprocessing: Quality control is performed on the collected raw data to remove outliers and erroneous data, a unified time and space reference system is established, a unified coordinate system is established, geometric, radiometric and atmospheric corrections are performed on remote sensing images, interpolation processing is performed on topographic data to generate a continuous digital elevation model, and time series analysis is performed on marine dynamic environment data to extract feature parameters.

[0243] Through the above-mentioned multi-source data collection and preprocessing, a complete basic database was established for the Shidaowan coastal zone protection and restoration project, providing data support for subsequent shoreline evolution analysis and restoration decisions.

[0244] S200, Analysis and Model Building of Beach Evolution Mechanism: In the Shidao Bay Coastal Zone Protection and Restoration Project, the specific implementation steps for the analysis and model building of beach evolution mechanism are as follows:

[0245] Analysis of shoreline evolution history: Based on multi-temporal remote sensing images, the changes in the location of the Shidao Bay shoreline between 1990 and 2020 were extracted, the rate of shoreline change was calculated, erosion and siltation sections were identified, the spatiotemporal distribution characteristics of shoreline changes were analyzed, and key change areas were determined.

[0246] Identification and quantification of influencing factors: Analyze the impact of natural dynamic factors such as waves, tides, and ocean currents on shoreline evolution, assess the impact of extreme events such as typhoons and storm surges on shoreline morphology, quantify the interference of human activities (such as land reclamation, port construction, and protective engineering) on ​​shoreline evolution, and establish a correlation model between influencing factors and shoreline evolution.

[0247] Analysis of sediment transport process: Based on field observation data, this study analyzes the sediment transport characteristics of Shidao Bay, identifies the main sediment sources and sinks, determines sediment transport paths, calculates sediment balance, assesses regional sediment supply and demand, and analyzes the interference mechanisms of human activities on sediment transport process.

[0248] Hydrodynamic numerical model construction: A two-dimensional tidal current numerical model of Shidao Bay was established, a wave propagation model was constructed, the wave field distribution was simulated, the storm surge process under extreme weather conditions was simulated, and the model was calibrated and verified using measured data;

[0249] Numerical model construction for shoreline evolution: Based on the hydrodynamic model, a numerical model for sediment transport is established, a shoreline topography evolution model is constructed, the process of shoreline change is simulated, a shoreline change prediction model is developed, the future shoreline location is predicted, and the model is validated and optimized using historical data.

[0250] Multi-scenario simulation and analysis: Design different sea-level rise scenarios to simulate the impact of climate change on shoreline evolution, simulate the impact of typhoons and storm surges of different intensities on shorelines, assess the impact of different human activity plans on shoreline stability, and analyze the response characteristics and evolution trends of shorelines under various scenarios.

[0251] Through the above-mentioned analysis and modeling of the shoreline evolution mechanism, the Shidaowan Coastal Zone Protection and Restoration Project has revealed the intrinsic mechanism of shoreline evolution in the region, established a reliable prediction model, and provided a theoretical basis and technical support for making scientific restoration decisions.

[0252] S300, Beach Restoration Scheme Design and Evaluation: In the Shidao Bay Coastal Zone Protection and Restoration Project, the specific implementation steps for beach restoration scheme design and evaluation are as follows:

[0253] Restoration goals were defined: Based on the analysis of shoreline evolution, key restoration areas in Shidao Bay were identified, and multi-dimensional restoration goals were set for shoreline stability, ecological function, and disaster prevention and mitigation. Short-term (5 years), medium-term (10 years), and long-term (20 years) phased restoration goals were formulated, and the expectations and requirements of various stakeholders for the restoration effect were clarified.

[0254] Restoration technology screening: Collect domestic and international cases and best practices of beach restoration technologies, evaluate the applicability of various restoration technologies in Shidao Bay, and screen the combination of restoration technologies suitable for the characteristics of Shidao Bay, including: beach replenishment and maintenance technology, ecological revetment technology, artificial reef construction technology, vegetation restoration technology and sediment interception and regulation technology.

[0255] Repair plan design:

[0256] Design Option 1: Primarily based on engineering measures, including the construction of offshore breakwaters, groynes systems, and beach replenishment;

[0257] Design Option 2: Focus on ecological restoration, including artificial reefs, ecological revetments, and vegetation restoration;

[0258] Design Scheme 3: A comprehensive scheme combining engineering measures and ecological restoration;

[0259] The engineering layout, scale, materials, and construction techniques of each plan should be refined.

[0260] Simulation and prediction of the effects of the schemes: Using the numerical model constructed by S200, the shoreline response after the implementation of each restoration scheme is simulated, the shoreline change trend and beach evolution process under different schemes are predicted, the protective effect of each scheme under extreme weather conditions is evaluated, and the impact of each scheme on the regional sediment transport pattern is analyzed.

[0261] The simulation results of the proposed solution are shown in Table 2.

[0262] Table 2. Predicted Degree of Achievement of Phased Goals for Shidao Bay Beach Restoration

[0263]

[0264] Multi-criteria comprehensive evaluation: Establish an evaluation index system that includes factors such as engineering effects, ecological impact, economic costs, and social benefits. Use the analytic hierarchy process (AHP) to determine the weight of each evaluation index, conduct multi-criteria comprehensive scoring on each remediation plan, analyze the advantages and disadvantages of each plan, and identify key limiting factors.

[0265] The evaluation results of the repair plan are shown in Table 3:

[0266] Table 3 Comparative Evaluation Table of Shidao Bay Beach Restoration Plans

[0267]

[0268] Optimal solution determination and optimization: Based on the comprehensive evaluation results, the optimal solution for the restoration of Shidao Bay beach is determined. The optimal solution is then optimized and adjusted locally to improve the implementation effect. A phased implementation plan and investment budget are formulated, and a restoration effect monitoring and evaluation scheme is designed.

[0269] Through the design and evaluation of the above-mentioned shoreline restoration scheme, the Shidao Bay Coastal Zone Protection and Restoration Project finally determined a comprehensive restoration scheme that combines engineering measures with ecological restoration. This scheme ensures shoreline stability while focusing on the restoration of ecosystem functions, and has good economic feasibility and social acceptance, providing scientific and technical support for the sustainable development of the Shidao Bay coastal zone.

[0270] The embodiments of the present invention have been described above, but the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention, all of which are within the protection scope of the present invention.

Claims

1. A beach restoration decision-making method based on multi-source data, characterized in that, Includes the following steps: S100 acquires multi-source data from beach monitoring, performs data fusion processing, and generates a feature tensor with unified spatiotemporal resolution. The multi-source data includes remote sensing imagery data, acoustic bathymetry data, GPS shoreline measurement data, wave buoy monitoring data, and tide gauge station records. S200, based on symbolic regression analysis using characteristic tensors, automatically discovers the governing equations of beach evolution; S300 inputs the evolution equation into the decision optimization framework to generate beach restoration decision schemes; The decision-making scheme for shoreline restoration includes the amount of artificial sand replenishment, the length of the sand-retaining dike, the elevation of the submerged dike, the number of groynes, and the inclination angle of the groynes; In S300, the specific steps include: S310, Repair Engineering Parameter Setting: Determine the set of control variables, set constraints, establish parameter association constraints, and generate an initial solution set that satisfies the constraints; S320, Evolutionary Prediction Simulation: Substitute the parameters of the repair scheme into the evolution equation; S330, Construction of multi-objective optimization function: Combining the definition of beach stability index, engineering cost function and ecological impact index to construct a comprehensive evaluation function; S340, Genetic Algorithm Optimization Solution: The repair scheme is represented by real number encoding, the algorithm parameters are set, simulated binary crossover and polynomial mutation are performed, and fast non-dominated sorting is adopted; S350, Solution Evaluation and Screening: Analyze the characteristics of the optimal solution set, evaluate the feasibility of the project, calculate the cost-benefit ratio, and consider decision preferences; In S350, analyzing the characteristics of the optimal solution set includes calculating the uniformity of solution distribution, evaluating the diversity index of solutions, and identifying key parameter combination patterns; evaluating engineering feasibility includes construction difficulty index, material availability, and environmental adaptability; calculating cost-benefit ratio includes stability improvement, engineering cost, and maximum allowable cost, and selecting the optimal solution. S360, Uncertainty Analysis: Using the Monte Carlo method, random samples are generated, evolution results are calculated, statistical confidence intervals are obtained, and robustness indices of the scheme are calculated; S370, Remediation Target Identification: Calculate the rate of shoreline change, identify erosion and sedimentation zones, assess topographic stability, and determine control sections; S380, Optimization of Intervention Timing: Analyze seasonal changes, identify state transition points, design and implement solutions based on state prediction models, allocate construction timelines for phased projects, and optimize project progress; S390, Decision Recommendation Generation: Determine project combinations, formulate layout plans, plan implementation sequence, compile construction guidelines, and output decision reports.

2. The beach restoration decision-making method based on multi-source data according to claim 1, characterized in that, S100 specifically includes the following steps: S110, multi-source data acquisition and preprocessing, acquiring remote sensing image data, including visible light and infrared band images, performing atmospheric and geometric correction, acquiring acoustic bathymetry data, performing noise filtering and depth calibration, acquiring GPS shoreline measurement data, performing coordinate transformation and outlier removal, acquiring wave buoy monitoring data, including quality control of wave height, period and direction data, acquiring tide gauge station record data, performing tide datum unification and data supplementation; S120, spatiotemporal alignment processing, uses bilinear interpolation to resample spatial data to a 5-meter grid resolution, uses cubic spline interpolation to resample time series to an hourly time scale, establishes a unified spatiotemporal reference system, ensures that all data sources are aligned in the same coordinate system, performs data synchronization checks, and ensures that timestamps from different data sources correspond accurately. S130, physical characteristic calculation, uses the finite difference method to calculate the spatial gradient of topographic elevation, uses the central difference scheme to calculate the time derivative of wave height, calculates the divergence of sediment flux based on the finite volume method, and constructs a composite physical quantity of wave-tide coupling term and wave steepness-water depth ratio. S140, Feature Tensor Generation: Design the data structure of the feature tensor, which includes spatial, temporal, and physical quantity dimensions. Fill the corresponding positions of the tensor with the aligned original observation data, integrate the calculated physical derivative features into the tensor, and perform data integrity checks and outlier handling.

3. The beach restoration decision-making method based on multi-source data according to claim 1, characterized in that, S200 specifically includes the following steps: S210, Data preprocessing and initialization, standardizes the feature tensor to unify the dimensions of each physical quantity, sets the candidate function set, initializes the genetic programming algorithm parameters, and constructs the fitness function; S220, a genetic programming evolutionary process, uses a tree-structured encoding method to represent candidate equations and performs crossover and mutation operations; S230, Sparse Regression Optimization, transforms the candidate equation terms generated by genetic programming into a coefficient matrix and applies the LASSO algorithm for sparsification. S240, physical consistency verification, checking the dimensional consistency of the equations, verifying the law of conservation of sediment mass, and verifying the law of conservation of wave energy; S250, time-varying feature analysis, uses a sliding window to segment the data sequence, identifies seasonal changes and extreme event periods, performs independent symbolic regression on each time period, generates a segmented evolution equation, and establishes a smooth transition mechanism between time periods; S260, Causal Structure Identification, uses the PC algorithm for causal discovery, calculates the structural edit distance of the DAG between adjacent periods, applies the PELT algorithm to detect structural abrupt points, constructs variable combination rules based on the DAG, and updates the candidate set of equation terms; S270, Coupling Effect Analysis, calculates the cross-correlation function of variable pairs, generates first-order and second-order coupling terms, and constructs composite physical quantities.

4. The beach restoration decision-making method based on multi-source data according to claim 3, characterized in that, The candidate function set includes: basic arithmetic operations; trigonometric functions; exponential functions; and power functions.

5. The beach restoration decision-making method based on multi-source data according to claim 4, characterized in that, Genetic programming algorithm parameters include population size, number of iterations, crossover probability, and mutation probability.

6. The beach restoration decision-making method based on multi-source data according to claim 5, characterized in that, In S220, the crossover operation includes randomly selecting two parent trees, randomly selecting the crossover point, and exchanging subtrees to generate a new individual.

7. The beach restoration decision-making method based on multi-source data according to claim 6, characterized in that, In S220, the mutation operation involves randomly selecting mutation points and replacing them with new operators / variables with probability, using a tournament selection method for iterative evolution.

8. A beach restoration decision-making system based on multi-source data, characterized in that, Performing the steps in the beach restoration decision-making method based on multi-source data as described in any one of claims 1-7 includes: Data acquisition and preprocessing module: responsible for acquiring and preprocessing multi-source monitoring data, performing spatiotemporal alignment processing on the data, unifying it to a 5-meter grid resolution and an hourly time scale, calculating physical features and generating feature tensors containing multi-scale physical process information; Evolutionary equation discovery module: Based on feature tensors, symbolic regression analysis is performed to automatically discover the governing equations of beach evolution through genetic programming algorithms, sparse regression optimization and physical consistency verification are performed, time-varying features and causal structures are analyzed, key coupling effects are identified, and finally evolutionary equations that can describe complex physical processes are generated. Decision optimization module: Input the evolution equation into the decision optimization framework, set the restoration project parameters and perform evolution prediction simulation, construct a multi-objective optimization function, solve the optimal solution through a genetic algorithm, perform scheme evaluation, uncertainty analysis and restoration target identification, optimize the timing of intervention, and finally generate a complete beach restoration decision scheme.

Images