Anti-floating water level prediction system and method

Abstract

The application relates to the technical field of geotechnical engineering, and discloses an anti-floating flood control water level prediction system and method, which comprises a data acquisition module, a data fusion processing module, a feature extraction module, a trend prediction module, a model training module, a prediction analysis module, an early warning evaluation module and a visual display module. When the anti-floating flood control water level prediction is carried out, the multi-source data fusion processing is established, the geological hydrological data, the meteorological data and the real-time monitoring data are standardized and intelligently fused, the influence of sensor precision, transmission interference and environmental noise can be eliminated, the consistency of multi-source heterogeneous data is ensured, high-quality anti-floating flood control water level prediction fusion data is provided for water level prediction, the prediction error caused by data quality problems is reduced, and the accuracy and reliability of the anti-floating flood control water level prediction result are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geotechnical engineering, in particular to an anti-floating design water level prediction system and method. BACKGROUND

[0002] The anti-floating design water level is generally divided into temporary engineering anti-floating design water level and permanent engineering anti-floating design water level, and the anti-floating design water level is a core parameter in building engineering design for resisting the buoyancy of groundwater, which is defined as the highest water level in history or the highest water level in the prediction period. The parameter directly affects the division of the waterproof grade of underground engineering, and its determination needs to consider hydrogeological conditions, historical data and prediction analysis.

[0003] At present, in the process of predicting the groundwater level, the prediction model depends on historical and real-time monitoring data, and cannot detect whether data anomalies caused by sensor accuracy, transmission interference and environmental noise occur in the data collection process in real time. When there are systematic bias and accidental error in the data source, the error of the prediction result is large, and the accuracy of the water level prediction cannot be guaranteed.

[0004] Therefore, the present application provides an anti-floating design water level prediction system and method to solve the above problems. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides an anti-floating design water level prediction system and method, which solves the problem of large prediction error and low accuracy of water level prediction in the background art.

[0006] To achieve the above purpose, the present application provides the following technical scheme: An anti-floating design water level prediction system and method, the method comprising the following steps:

[0007] S1, collecting geological hydrological data, meteorological data and real-time monitoring data of the target area to generate anti-floating design water level prediction original data set;

[0008] S2, performing multi-source data fusion processing on the anti-floating design water level prediction original data set to generate anti-floating design water level prediction fusion data;

[0009] S3, performing geological hydrological feature parameter extraction processing based on the anti-floating design water level prediction fusion data to generate a geological hydrological feature parameter data set;

[0010] S4, performing dynamic change trend analysis of the groundwater level based on the geological hydrological feature parameter data set to generate dynamic change trend prediction data of the groundwater level;

[0011] S5, based on the groundwater level dynamic change trend prediction data combining artificial intelligence algorithm for anti-floating water level prediction model training processing, generating anti-floating water level prediction model;

[0012] S6, using the anti-floating water level prediction model for real-time anti-floating water level prediction analysis, generating anti-floating water level prediction result data;

[0013] S7, according to the anti-floating water level prediction result data for anti-floating warning level evaluation processing, generating anti-floating warning level data;

[0014] S8, based on the anti-floating warning level data to generate anti-floating water level prediction report and visual display.

[0015] Preferably, the S1 in collecting geological and hydrological data, meteorological data and real-time monitoring data of target area includes the following steps:

[0016] S11, collecting geological structure parameters, rock and soil mechanics parameters and groundwater parameters of target area by geological exploration equipment, generating geological and hydrological data;

[0017] S12, collecting precipitation, evaporation, temperature change parameters of target area by meteorological monitoring equipment, generating meteorological data;

[0018] S13, collecting real-time water level change data, pore water pressure data and soil moisture content data of target area by water level monitoring sensor network, generating real-time monitoring data;

[0019] S14, the geological and hydrological data, meteorological data and real-time monitoring data are standardized, generating anti-floating water level prediction original data set.

[0020] Preferably, the S2 in the multi-source data fusion processing includes the following steps:

[0021] S21, obtaining the anti-floating water level prediction original data set;

[0022] S22, using wavelet transform algorithm to process time-frequency feature alignment of multi-source heterogeneous data, generating time-frequency alignment data;

[0023] S23, based on Kalman filter algorithm for data fusion processing of time-frequency alignment data, generating anti-floating water level prediction fusion data.

[0024] Preferably, the S3 in the geological and hydrological feature parameter extraction processing includes the following steps:

[0025] S31, extracting the geological permeability coefficient, water storage coefficient and water supply degree parameter in the anti-floating water level prediction fusion data;

[0026] S32. Calculate the groundwater runoff intensity, recharge and discharge parameters based on numerical analysis methods;

[0027] S33. Generate a dataset of geological and hydrological feature parameters containing vectors of geological and hydrological feature parameters.

[0028] Preferably, the analysis of the dynamic change trend of groundwater level in step S4 specifically includes the following steps:

[0029] S41. Establish a differential equation model for the dynamic change of groundwater level:

[0030] ;

[0031] in The water storage coefficient, Water level height Permeability coefficient, For source and sink items;

[0032] S42. Solve the differential equation for the dynamic change of groundwater level using the finite difference method;

[0033] S43. Generate dynamic trend forecast data containing future time series water level prediction values.

[0034] Preferably, the training process for the anti-buoyancy design water level prediction model in step S5 specifically includes the following steps:

[0035] S51. Construct a deep learning-based neural network model for predicting anti-buoyancy design water levels;

[0036] S52. Use the geological and hydrological characteristic parameter dataset and the groundwater level dynamic change trend prediction data as training samples;

[0037] S53. Optimize the neural network weight parameters using the backpropagation algorithm;

[0038] S54. When the model prediction accuracy reaches the preset threshold, generate an anti-buoyancy design water level prediction model.

[0039] Preferably, the training process for the anti-buoyancy design water level prediction model in S5 further includes model optimization and dynamic update operations:

[0040] S55. Establish a continuous monitoring mechanism for the performance of the prediction model, and collect error data between the predicted water level of the anti-buoyancy design and the actual water level monitoring data in real time.

[0041] S56. Construct a model error trend evaluation function based on time series analysis. When the continuous error of the evaluation function exceeds the preset tolerance threshold, the model retraining instruction is automatically triggered.

[0042] S57. Call the newly generated geological and hydrological characteristic parameter dataset and the updated real-time monitoring data, and use the incremental learning algorithm to fine-tune the parameters and optimize the structure of the anti-buoyancy design water level prediction model.

[0043] S58. Compare and verify the optimized model performance indicators with historical versions, generate a model iteration update log, and update the anti-buoyancy design water level prediction model.

[0044] Preferably, the real-time anti-buoyancy design water level prediction analysis in step S6 specifically includes the following steps:

[0045] S61. Obtain real-time monitoring data and input it into the anti-buoyancy design water level prediction model;

[0046] S62. Generate predicted values ​​of anti-buoyancy design water level for a specified future time period through model calculation;

[0047] S63. Generate anti-buoyancy design water level prediction results data including predicted water level value, confidence interval and prediction accuracy.

[0048] Preferably, the anti-buoyancy defense early warning level assessment process in S7 specifically includes the following steps:

[0049] S71. Establish an evaluation index system for anti-buoyancy defense early warning levels;

[0050] S72. Determine the warning level based on the difference between the predicted water level and the designed flood control level;

[0051] S73. Generate anti-buoyancy defense early warning level data that includes early warning level, risk level and recommended measures.

[0052] Preferably, the system includes:

[0053] The data acquisition module receives raw signals from external exploration, monitoring and meteorological equipment, and outputs raw datasets for anti-buoyancy design water level prediction after standardization processing.

[0054] The data fusion processing module receives the original dataset of the anti-buoyancy design water level prediction, processes it through a multi-source data fusion algorithm, and outputs the anti-buoyancy design water level prediction fusion data.

[0055] The feature extraction module receives the fused data of the anti-buoyancy design water level prediction and outputs the geological and hydrological feature parameter dataset after parsing the geological and hydrological feature parameters.

[0056] The trend prediction module receives the geological and hydrological characteristic parameter dataset, and outputs groundwater level dynamic change trend prediction data after analyzing the dynamic change trend of groundwater level.

[0057] The model training module receives the predicted data on the dynamic change trend of groundwater level and the dataset of geological and hydrological characteristic parameters, and outputs the anti-buoyancy design flood level prediction model after training and optimization by artificial intelligence algorithms.

[0058] The predictive analysis module receives real-time monitoring data and the anti-buoyancy design water level prediction model, and outputs the anti-buoyancy design water level prediction result data after real-time water level prediction calculation.

[0059] The early warning assessment module receives the predicted results of the anti-buoyancy design water level, processes them through the early warning level assessment logic, and outputs the anti-buoyancy design early warning level data.

[0060] The visualization module receives the anti-buoyancy fortification early warning level data and prediction result data, and outputs an anti-buoyancy fortification water level prediction report after report generation and visualization rendering. Beneficial effects

[0061] Compared with the prior art, the present invention provides a system and method for predicting the water level of an anti-buoyancy structure, which has the following beneficial effects:

[0062] 1. In this invention, when the system performs anti-buoyancy design water level prediction, it establishes multi-source data fusion processing to standardize and intelligently fuse geological and hydrological data, meteorological data, and real-time monitoring data. This eliminates the influence of sensor accuracy, transmission interference, and environmental noise, ensures the consistency of multi-source heterogeneous data, provides high-quality anti-buoyancy design water level prediction fusion data for water level prediction, reduces prediction errors caused by data quality issues, and improves the accuracy and reliability of anti-buoyancy design water level prediction results.

[0063] 2. In this invention, when training and updating the anti-buoyancy design water level prediction model, the system constructs a continuous monitoring mechanism and a dynamic optimization mechanism for the prediction model performance. It tracks the trend of prediction error changes in real time and automatically triggers model retraining instructions, enabling the anti-buoyancy design water level prediction model to have adaptive adjustment capabilities. This allows it to cope with changes in data feature distribution and model performance degradation, avoids the occurrence of prediction model failure, ensures the long-term stability and generalization ability of the water level prediction system, and enhances the reliability of anti-buoyancy design decisions.

[0064] 3. In this invention, when the system conducts an anti-buoyancy fortification early warning level assessment, it establishes a quantitative and unified early warning level mapping table and an adaptive threshold matching mechanism. Based on different engineering geological conditions and hydrological environmental characteristics, it intelligently matches the early warning threshold, enabling accurate early warnings for multiple regions and levels. This avoids situations where the early warning level judgment is too strict or too lenient, ensuring the matching degree between the anti-buoyancy fortification early warning level data and the actual risk, and improving the accuracy of anti-buoyancy fortification early warning and the engineering guidance effect. Attached Figure Description

[0065] Fig. 1This is a flowchart of a method for predicting the water level of an anti-buoyancy structure according to the present invention;

[0066] Fig. 2 This is a schematic diagram of the anti-buoyancy water level prediction system of the present invention. Detailed Implementation

[0067] The technical solutions of 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.

[0068] For specific implementation examples, please refer to: Figs. 1-2 The present invention relates to a system and method for predicting the water level of an anti-buoyancy structure, the method comprising the following steps:

[0069] S1. Collect geological and hydrological data, meteorological data and real-time monitoring data of the target area to generate the original dataset for predicting the anti-buoyancy design water level;

[0070] S2. Perform multi-source data fusion processing on the original dataset for anti-floating design water level prediction to generate anti-floating design water level prediction fused data.

[0071] S3. Based on the fusion data of anti-buoyancy design water level prediction, geological and hydrological characteristic parameters are extracted and processed to generate a geological and hydrological characteristic parameter dataset.

[0072] S4. Analyze the dynamic change trend of groundwater level based on the geological and hydrological characteristic parameter dataset, and generate groundwater level dynamic change trend prediction data.

[0073] S5. Based on the prediction data of the dynamic change trend of groundwater level, combined with artificial intelligence algorithm, the anti-buoyancy design water level prediction model is trained and processed to generate the anti-buoyancy design water level prediction model.

[0074] S6. Use the anti-buoyancy design water level prediction model to perform real-time anti-buoyancy design water level prediction analysis and generate anti-buoyancy design water level prediction result data.

[0075] S7. Based on the predicted data of the anti-buoyancy design water level, conduct an assessment of the anti-buoyancy design early warning level and generate anti-buoyancy design early warning level data.

[0076] S8. Generate a prediction report of the anti-buoyancy design water level based on the anti-buoyancy design early warning level data and display it visually.

[0077] The specific steps involved in collecting geological, hydrological, meteorological, and real-time monitoring data of the target area in S1 are as follows:

[0078] S11. Collect geological structure parameters, soil and rock mechanics parameters and groundwater parameters of the target area using geological exploration equipment to generate geological and hydrological data.

[0079] S12. Collect precipitation, evaporation, and temperature change parameters of the target area using meteorological monitoring equipment, and generate meteorological data;

[0080] S13. Collect real-time water level change data, pore water pressure data and soil moisture content data of the target area through a water level monitoring sensor network, and generate real-time monitoring data.

[0081] S14. Standardize the geological and hydrological data, meteorological data, and real-time monitoring data to generate the original dataset for predicting the anti-buoyancy design water level.

[0082] The multi-source data fusion processing in S2 specifically includes the following steps:

[0083] S21. Obtain the original dataset for predicting the anti-buoyancy design water level;

[0084] S22. Use wavelet transform algorithm to perform time-frequency feature alignment processing on multi-source heterogeneous data to generate time-frequency aligned data, including the following steps:

[0085] S221. Select the Db4 wavelet as the basis function and perform multi-scale decomposition on each time series data in the original dataset for anti-floating flood level prediction to obtain the high-frequency coefficients and low-frequency coefficients of each data sequence.

[0086] S222. For the high-frequency coefficients after decomposition, a soft thresholding function is used for noise reduction to eliminate random interference during data acquisition. The soft thresholding function is defined as follows:

[0087] ;

[0088] in This represents the processed coefficient value. The input high-frequency wavelet coefficients, For threshold parameters, For symbolic functions, This is a function to find the maximum value.

[0089] S223. Using the timestamps of geological and hydrological data as a benchmark, perform time-domain translation transformation on the coefficients of each data sequence after noise reduction to ensure that all data have consistent physical meaning at the same point in time.

[0090] S224. Reconstruct the wavelet coefficients of each sequence after time-domain alignment to generate time-frequency feature aligned data;

[0091] S23. Based on the Kalman filter algorithm, perform data fusion processing on the time-frequency aligned data to generate anti-buoyancy design water level prediction fused data, including the following steps:

[0092] S231. Establish the system state equations and observation equations with groundwater level as the core state variable, and establish the state-space model of the discrete-time system as follows:

[0093] System state equations:

[0094] ;

[0095] in for The system state vector at time t. Here is the state transition matrix. To control the input matrix, To control the input vector, This is the process noise vector;

[0096] Observation equation:

[0097] ;

[0098] in for The observation vector at time t, For the observation matrix, For the observed noise vector;

[0099] S232. Initialize the state vector estimate and error covariance matrix of the Kalman filter;

[0100] S233. Perform Kalman gain calculation, and use real-time monitoring data to optimally correct the state prediction value, then calculate the Kalman gain matrix. :

[0101] ;

[0102] in Here is the Kalman gain matrix. Let be the prior error covariance matrix. Let be the posterior error covariance matrix. To observe the noise covariance matrix, For the observation model matrix, For the observation model matrix transpose, For time steps;

[0103] S234. Update the error covariance matrix and iteratively execute the prediction and correction steps until all time-frequency aligned data has been processed, and output the anti-buoyancy design water level prediction fusion data.

[0104] The extraction and processing of geological and hydrological characteristic parameters in S3 specifically includes the following steps:

[0105] S31. Extract the geological permeability coefficient, water storage coefficient, and specific yield parameters from the fusion data of the anti-buoyancy design water level prediction.

[0106] S32. Calculate the groundwater runoff intensity, recharge and discharge parameters based on numerical analysis methods;

[0107] S33. Generate a dataset of geological and hydrological feature parameters containing vectors of geological and hydrological feature parameters.

[0108] The specific steps involved in analyzing the dynamic trend of groundwater level changes in S4 are as follows:

[0109] S41. Establish a differential equation model for the dynamic change of groundwater level:

[0110] ;

[0111] in The water storage coefficient, Water level height Permeability coefficient, For source and sink items;

[0112] S42. Solve the differential equation for the dynamic change of groundwater level using the finite difference method, including the following steps:

[0113] S421. Perform three-dimensional meshing on the target region, discretizing the continuous solution domain into a finite number of mesh elements;

[0114] S422. Approximate the partial derivative terms in the differential equation using the difference quotient form, construct a system of linear equations corresponding to each grid node, and approximate the partial derivative terms in the differential equation using the difference quotient form:

[0115] The time partial derivatives are obtained using forward differencing:

[0116] ;

[0117] in The rate of change of water level over time. This represents the water level at a future time. These are the coordinates of the three-dimensional spatial grid nodes. For time step index, This represents the current water level. Index for the current time step;

[0118] Spatial partial derivatives are expressed using the central difference scheme:

[0119] ;

[0120] in Permeability coefficient, For water level gradient, Water level height For spatial step size, For dimension indexing;

[0121] S423. Introduce boundary conditions and initial conditions determined by the geological and hydrological characteristic parameter dataset;

[0122] S424. The preprocessing conjugate gradient method is used to iteratively solve the linear equation system to obtain the water level prediction value of each grid node at each discrete time step, thereby generating groundwater level dynamic change trend prediction data.

[0123] S43. Generate dynamic trend forecast data containing future time series water level prediction values.

[0124] The training process for the anti-buoyancy design water level prediction model in S5 includes the following steps:

[0125] S51. Construct a deep learning-based neural network model for predicting anti-buoyancy design water levels. The model construction includes the following steps:

[0126] S511. The neural network model adopts a multi-input branch fusion structure, including geological and hydrological feature input branches, dynamic trend input branches, and real-time monitoring data input branches;

[0127] S512, the geological and hydrological feature input branch receives the geological and hydrological feature parameter dataset and performs feature extraction through two fully connected layers;

[0128] S513: The dynamic trend input branch receives the predicted data of the dynamic change trend of groundwater level, and extracts the temporal features through a one-dimensional temporal convolutional layer.

[0129] S514, The real-time monitoring data input branch receives real-time monitoring data, which is then processed by the data standardization layer and input into the model;

[0130] S515. The output features of each branch are concatenated in the feature fusion layer and then transformed nonlinearly through a three-layer fully connected network including a Dropout layer.

[0131] S516. The final output layer uses a linear activation function to output the predicted value of the anti-buoyancy design water level for a specified future time period.

[0132] S517. All hidden layers of the model use the ReLU activation function, and the weights are initialized using the He normal distribution initialization method.

[0133] S52. Geological and hydrological characteristic parameter datasets and groundwater level dynamic change trend prediction data are used as training samples;

[0134] S53. Optimize the neural network weight parameters using the backpropagation algorithm, wherein the backpropagation algorithm includes the following steps:

[0135] S531, Forward computation: Input the training samples into the neural network model, calculate layer by layer, and obtain the predicted value of the output layer;

[0136] S532. Loss Calculation: Calculate the error between the model's predicted value and the true value using the mean squared error function.

[0137] ;

[0138] in For loss function, The normalization coefficient is... For sample index, For the total sample size, For predicted values, The actual value;

[0139] S533, Backpropagation of Error: The error is propagated from the output layer to the input layer in the back and the gradient corresponding to each weight parameter is calculated using the chain rule.

[0140] S534, Parameter Update: An adaptive moment estimation algorithm is used to iteratively update the weight parameters in the neural network based on gradient information.

[0141] ;

[0142] in For the updated parameters, For the current parameter, For learning rate, This is the bias correction value for the first-order moment estimate. This is the square root of the bias correction value for the second-order moment estimate. It is a very small constant. For time steps;

[0143] S54. When the model prediction accuracy reaches the preset threshold, generate an anti-buoyancy design water level prediction model.

[0144] The training process for the anti-buoyancy design water level prediction model in S5 also includes model optimization and dynamic update operations:

[0145] S55. Establish a continuous monitoring mechanism for the performance of the prediction model, and collect error data between the predicted water level of the anti-buoyancy design flood control level and the actual water level monitoring data in real time. The establishment of the continuous monitoring mechanism for the performance of the prediction model includes the following steps:

[0146] S551. Real-time acquisition of predicted water level values ​​from anti-buoyancy design water level prediction data. Compared with actual water level monitoring data Construct a time series error dataset ;

[0147] S552. Calculate the absolute error value at each time point in the time series error dataset. With relative error value The calculation formulas are as follows:

[0148] ;

[0149] ;

[0150] in For a point in time;

[0151] S553, Set the model performance tolerance threshold and ,in The absolute error threshold. This is the relative error threshold;

[0152] S554, Continuously compare absolute error values With relative error value Do they all exceed the tolerance threshold at the same time? and Record the number of consecutive times the limit is exceeded. ;

[0153] S555, When the limit is exceeded consecutively Reaching the preset number of times When the model performance is determined to be degraded, a retraining trigger instruction is automatically generated and output.

[0154] S556, Transfer the timing error dataset Exceeding limits and retraining trigger commands are integrated into a model performance monitoring log and updated in real time;

[0155] S56. Construct a model error trend evaluation function based on time series analysis. When the continuous error of the evaluation function exceeds a preset tolerance threshold, automatically trigger the model retraining instruction, including the following steps:

[0156] S561. Obtain the timing error dataset The absolute error values ​​at multiple consecutive time points ;

[0157] S562. Calculate the moving average of the absolute error values ​​at multiple consecutive time points. This is used as the model error trend evaluation value, and its calculation formula is:

[0158] ;

[0159] in At the current time point, For the set calculation window size, For index variables;

[0160] S563, evaluate the model error trend value Compared with the preset absolute error tolerance threshold Compare;

[0161] S564, when continuous All of these calculations satisfy the following conditions. When needed, it automatically generates and outputs the model retraining trigger command;

[0162] S57. Using the newly generated geological and hydrological characteristic parameter dataset and updated real-time monitoring data, an incremental learning algorithm is employed to fine-tune the parameters and optimize the structure of the adversarial floating flood control level prediction model, including the following steps:

[0163] S571, Weight parameters of the feature extraction layer in the prediction model for frozen anti-buoyancy design water level;

[0164] S572. Add a new fully connected adaptation layer before the model output layer to learn new data features;

[0165] S573. Fine-tune the weights of the adaptation layer and the last two layers of the model using new training data with a smaller learning rate.

[0166] S574. Evaluate the performance of the fine-tuned model on the validation set. If the performance improves, retain the update; otherwise, roll back to the previous version, thereby achieving dynamic optimization of the model structure.

[0167] S58. Compare and verify the optimized model performance indicators with historical versions, generate model iteration update logs, and update the anti-buoyancy design water level prediction model.

[0168] The real-time anti-buoyancy design water level prediction analysis in S6 specifically includes the following steps:

[0169] S61. Obtain real-time monitoring data and input it into the anti-buoyancy design water level prediction model;

[0170] S62. Generate predicted values ​​of the anti-buoyancy design water level for a specified future time period through model calculation, including the following steps:

[0171] S621. Obtain real-time monitoring data at the current moment. And compared it with the geological and hydrological characteristic parameter dataset Combine them to construct the model input vector. ;

[0172] S622, Input the model vector Input into the trained anti-buoyancy design water level prediction model;

[0173] S623, the anti-buoyancy design water level prediction model calculates based on its internal neural network weight parameters and outputs the future water level for a specified time period. The sequence of predicted water levels at various time points within the region:

[0174] ;

[0175] in The preset future prediction step size;

[0176] S624, Sequence of predicted water levels Encapsulate the data with the corresponding timestamp to generate the anti-buoyancy design water level prediction result data;

[0177] S63. Generate anti-buoyancy design water level prediction results data including predicted water level value, confidence interval and prediction accuracy.

[0178] The specific steps involved in the anti-buoyancy defense early warning level assessment in S7 are as follows:

[0179] S71. Establish an evaluation index system for anti-buoyancy defense early warning levels;

[0180] S72. Determine the warning level based on the difference between the predicted water level and the flood control level, including the following steps:

[0181] S721. Calculate the predicted water level at a specific future time point. Water level in engineering design absolute difference between The calculation formula is:

[0182] ;

[0183] S722. Query the preset warning level mapping table, which defines different difference ranges. The corresponding warning levels, and the warning levels and their corresponding relationships, include:

[0184] When 0≤ When the value is less than 0.5, it corresponds to a Level IV blue alert, and the normal monitoring frequency should be maintained.

[0185] When 0.5≤ When the value is less than 1.0, a Level 3 Yellow Alert is issued, and the monitoring frequency is increased and patrols are initiated.

[0186] When 1.0≤ When the value is less than 2.0, a Level II orange alert is issued, monitoring is upgraded to real-time status, and emergency drainage is prepared.

[0187] when When the value is ≥2.0, it corresponds to a Level 1 Red Alert, and the emergency response plan should be activated immediately and drainage operations should be carried out.

[0188] S723, when When the level 1 warning threshold is exceeded, a warning information generation command is automatically triggered.

[0189] S724. Associate the early warning level with the corresponding risk level and recommended measures to generate anti-buoyancy defense early warning level data;

[0190] S73. Generate anti-buoyancy defense early warning level data that includes early warning level, risk level and recommended measures.

[0191] The system includes:

[0192] The data acquisition module receives raw signals from external exploration, monitoring and meteorological equipment, and outputs raw datasets for anti-buoyancy design water level prediction after standardization processing.

[0193] The data fusion processing module receives the original dataset of anti-buoyancy design water level prediction, processes it through a multi-source data fusion algorithm, and outputs the anti-buoyancy design water level prediction fusion data. The multi-source data fusion algorithm is configured to perform the following: time-frequency feature alignment is achieved through a wavelet transform unit, and then optimal data fusion is performed through a Kalman filter unit, finally outputting the anti-buoyancy design water level prediction fusion data.

[0194] The feature extraction module receives the fusion data of the anti-buoyancy design water level prediction, and outputs the geological and hydrological feature parameter dataset after parsing the geological and hydrological feature parameters.

[0195] The trend prediction module receives a dataset of geological and hydrological characteristic parameters, and outputs groundwater level dynamic change trend prediction data after groundwater level dynamic change trend analysis. The groundwater level dynamic change trend analysis is configured to perform: numerically solve the discretized groundwater motion differential equation based on the finite difference solver, and output future time series water level dynamic change trend prediction data.

[0196] The model training module receives groundwater level dynamic change trend prediction data and geological and hydrological characteristic parameter datasets. After training and optimization by artificial intelligence algorithms, it outputs an anti-buoyancy design flood level prediction model. The artificial intelligence algorithm is configured to: optimize the weight parameters of the neural network through backpropagation algorithm and integrate incremental learning units to achieve continuous fine-tuning and dynamic updating of the model.

[0197] The predictive analysis module receives real-time monitoring data and the anti-buoyancy design water level prediction model, and outputs the anti-buoyancy design water level prediction result data after real-time water level prediction calculation.

[0198] The early warning assessment module receives the predicted results of the anti-buoyancy design water level, and outputs the anti-buoyancy design early warning level data after processing by the early warning level assessment logic.

[0199] The visualization module receives anti-buoyancy fortification early warning level data and prediction results data, and outputs an anti-buoyancy fortification water level prediction report after report generation and visualization rendering.

[0200] The operation steps of this anti-buoyancy design water level prediction system and method are as follows:

[0201] Step 1: Multi-source data acquisition and standardization processing

[0202] The system utilizes a network of geological exploration equipment, meteorological monitoring equipment, and water level monitoring sensors deployed in the target area to collect geological and hydrological data on geological structure parameters, geotechnical mechanics parameters, and groundwater parameters; meteorological data such as precipitation, evaporation, and temperature changes; and real-time monitoring data on water level changes, pore water pressure, and soil moisture content. Subsequently, the system standardizes this multi-source heterogeneous data to generate a unified raw dataset for anti-buoyancy design flood level prediction, providing a standardized data foundation for subsequent data fusion and predictive analysis.

[0203] Step 2: Intelligent Data Fusion and Feature Extraction

[0204] The system employs wavelet transform algorithm to align the time-frequency features of multi-source heterogeneous data in the original dataset for anti-floating flood control level prediction, eliminating random interference. Then, based on Kalman filtering algorithm, it performs optimal fusion processing on the time-frequency aligned data to generate high-quality fused data for anti-floating flood control level prediction. Subsequently, the system extracts parameters such as geological permeability coefficient, storage coefficient, specific yield, groundwater runoff intensity, recharge, and discharge from the fused data, constructing a geological and hydrological characteristic parameter dataset, providing core input for trend analysis.

[0205] Step 3: Analysis and Prediction of Water Level Dynamics

[0206] Based on a dataset of geological and hydrological characteristic parameters, the system establishes a differential equation model for the dynamic change of groundwater level with groundwater level as the core state quantity, and uses the finite difference method for numerical solution to generate groundwater level dynamic change trend prediction data containing future time series water level prediction values.

[0207] Step 4: Training and Self-Optimization of the Intelligent Prediction Model

[0208] The system constructs a deep learning-based neural network model for predicting anti-buoyancy design water levels. Geological and hydrological characteristic parameter datasets and predicted data on dynamic changes in groundwater levels are used as training samples. The backpropagation algorithm is employed to optimize the neural network weight parameters, completing model training. A continuous performance monitoring mechanism for the prediction model is established, collecting real-time data on the error between prediction results and actual monitoring data. When performance degrades, a model retraining command is automatically triggered, and an incremental learning algorithm is used to fine-tune the model's parameters and optimize its structure.

[0209] Step 5: Real-time water level prediction and early warning assessment

[0210] The system inputs real-time monitoring data into the trained anti-buoyancy design flood level prediction model, calculates and generates predicted anti-buoyancy design flood levels and their confidence intervals for a specified future time period, forming complete anti-buoyancy design flood level prediction results data. Subsequently, the system calculates the absolute difference between the predicted water level and the engineering design flood level. It queries the preset warning level mapping table, automatically determines the warning level, and generates anti-buoyancy defense warning level data containing the warning level, risk degree, and recommended measures, thereby realizing the quantification and assessment of risk.

[0211] Step Six: Visualization and Decision Support

[0212] The system ultimately integrates the anti-buoyancy design water level prediction data with the anti-buoyancy design early warning level data, and generates an intuitive anti-buoyancy design water level prediction report through data visualization technology, forming a complete technical closed loop from data perception to intelligent decision-making.

[0213] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0214] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for predicting the water level of an anti-buoyancy design, characterized in that: The method includes the following steps: S1. Collect geological and hydrological data, meteorological data and real-time monitoring data of the target area to generate the original dataset for predicting the anti-buoyancy design water level; S2. Perform multi-source data fusion processing on the original dataset of the anti-buoyancy design water level prediction to generate anti-buoyancy design water level prediction fusion data. S3. Based on the predicted and fused data of the anti-buoyancy design water level, perform geological and hydrological feature parameter extraction and processing to generate a geological and hydrological feature parameter dataset, including: S31. Extract the geological permeability coefficient, water storage coefficient, and water yield parameters from the fusion data of the anti-buoyancy design water level prediction. S32. Calculate the groundwater runoff intensity, recharge and discharge parameters based on numerical analysis methods; S33. Generate a dataset of geological and hydrological feature parameters containing vectors of geological and hydrological feature parameters; S4. Based on the aforementioned geological and hydrological characteristic parameter dataset, perform dynamic trend analysis of groundwater level changes to generate predicted data for groundwater level dynamic trend changes, including: S41. Establish a differential equation model for the dynamic change of groundwater level: ； in The water storage coefficient, Water level height Permeability coefficient, For source and sink items; S42. Solve the differential equation for the dynamic change of groundwater level using the finite difference method; S43. Generate dynamic trend forecast data containing future time series water level prediction values; S5. Based on the predicted data of dynamic changes in groundwater level, combined with artificial intelligence algorithms, a prediction model for the anti-buoyancy design water level is trained to generate an anti-buoyancy design water level prediction model, including: S51. Construct a deep learning-based neural network model for predicting anti-buoyancy design water levels; S52. Use the geological and hydrological characteristic parameter dataset and the groundwater level dynamic change trend prediction data as training samples; S53. Optimize the neural network weight parameters using the backpropagation algorithm; S54. When the model prediction accuracy reaches the preset threshold, generate an anti-buoyancy design water level prediction model. S6. Use the anti-buoyancy design water level prediction model to perform real-time anti-buoyancy design water level prediction analysis and generate anti-buoyancy design water level prediction result data. S7. Based on the predicted anti-buoyancy design water level data, perform anti-buoyancy design early warning level assessment and generate anti-buoyancy design early warning level data. S8. Generate an anti-buoyancy design water level prediction report based on the anti-buoyancy design early warning level data and display it visually.

2. The method for predicting the anti-buoyancy design water level according to claim 1, characterized in that: The specific steps involved in collecting geological and hydrological data, meteorological data, and real-time monitoring data of the target area in S1 are as follows: S11. Collect geological structure parameters, soil and rock mechanics parameters and groundwater parameters of the target area using geological exploration equipment to generate geological and hydrological data. S12. Collect precipitation, evaporation, and temperature change parameters of the target area using meteorological monitoring equipment, and generate meteorological data; S13. Collect real-time water level change data, pore water pressure data and soil moisture content data of the target area through a water level monitoring sensor network, and generate real-time monitoring data. S14. Standardize the geological and hydrological data, meteorological data, and real-time monitoring data to generate a raw dataset for predicting the anti-buoyancy design water level.

3. The method for predicting the anti-buoyancy design water level according to claim 1, characterized in that: The multi-source data fusion processing in S2 specifically includes the following steps: S21. Obtain the original dataset of the anti-buoyancy design water level prediction; S22. Use wavelet transform algorithm to perform time-frequency feature alignment processing on multi-source heterogeneous data to generate time-frequency aligned data; S23. Based on the Kalman filter algorithm, perform data fusion processing on the time-frequency aligned data to generate anti-buoyancy design water level prediction fused data.

4. The method for predicting the anti-buoyancy design water level according to claim 1, characterized in that: The training process for the anti-buoyancy design water level prediction model in S5 also includes model optimization and dynamic update operations: S55. Establish a continuous monitoring mechanism for the performance of the prediction model, and collect error data between the predicted water level of the anti-buoyancy design and the actual water level monitoring data in real time. S56. Construct a model error trend evaluation function based on time series analysis. When the continuous error of the evaluation function exceeds the preset tolerance threshold, the model retraining instruction is automatically triggered. S57. Call the newly generated geological and hydrological characteristic parameter dataset and the updated real-time monitoring data, and use the incremental learning algorithm to fine-tune the parameters and optimize the structure of the anti-buoyancy design water level prediction model. S58. Compare and verify the optimized model performance indicators with historical versions, generate a model iteration update log, and update the anti-buoyancy design water level prediction model.

5. The method for predicting the anti-buoyancy design water level according to claim 1, characterized in that: The real-time anti-buoyancy design water level prediction analysis in S6 specifically includes the following steps: S61. Obtain real-time monitoring data and input it into the anti-buoyancy design water level prediction model; S62. Generate predicted values ​​of anti-buoyancy design water level for a specified future time period through model calculation; S63. Generate anti-buoyancy design water level prediction results data including predicted water level value, confidence interval and prediction accuracy.

6. The method for predicting the anti-buoyancy design water level according to claim 1, characterized in that: The process of assessing the anti-buoyancy defense early warning level in S7 specifically includes the following steps: S71. Establish an evaluation index system for anti-buoyancy defense early warning levels; S72. Determine the warning level based on the difference between the predicted water level and the designed flood control level; S73. Generate anti-buoyancy defense early warning level data that includes early warning level, risk level and recommended measures.

7. A system for predicting the level of an anti-buoyancy design site, used to implement the method for predicting the level of an anti-buoyancy design site according to any one of claims 1-6, characterized in that: The system includes: The data acquisition module receives raw signals from external exploration, monitoring and meteorological equipment, and outputs raw datasets for anti-buoyancy design water level prediction after standardization processing. The data fusion processing module receives the original dataset of the anti-buoyancy design water level prediction, processes it through a multi-source data fusion algorithm, and outputs the anti-buoyancy design water level prediction fusion data. The feature extraction module receives the fused data of the anti-buoyancy design water level prediction and outputs the geological and hydrological feature parameter dataset after parsing the geological and hydrological feature parameters. The trend prediction module receives the geological and hydrological characteristic parameter dataset, and outputs groundwater level dynamic change trend prediction data after analyzing the dynamic change trend of groundwater level. The model training module receives the predicted data on the dynamic change trend of groundwater level and the dataset of geological and hydrological characteristic parameters, and outputs the anti-buoyancy design flood level prediction model after training and optimization by artificial intelligence algorithms. The predictive analysis module receives real-time monitoring data and the anti-buoyancy design water level prediction model, and outputs the anti-buoyancy design water level prediction result data after real-time water level prediction calculation. The early warning assessment module receives the predicted results of the anti-buoyancy design water level, processes them through the early warning level assessment logic, and outputs the anti-buoyancy design early warning level data. The visualization module receives the anti-buoyancy fortification early warning level data and prediction result data, and outputs an anti-buoyancy fortification water level prediction report after report generation and visualization rendering.

Images