A digital warning method and system for bay hydrological conditions

Abstract

The present invention relates to the technical field of marine disaster warning, and discloses a digital bay hydrological condition warning method and system. The method includes collecting historical meteorological data and real-time meteorological data of the bay area to be predicted; according to the historical meteorological data, using deep reinforcement learning and genetic algorithm to establish a relationship model between wind and waves, and generating wave prediction data based on the relationship model and the real-time wind data of the bay area to be predicted; synthesizing a synthetic wind field of the bay area to be predicted based on the theoretical model wind field and the background wind field; according to the synthetic wind field and real-time meteorological data, using a hydrodynamic mathematical model and a tidal current mathematical model to conduct storm surge simulation; based on the wave prediction data and the storm surge simulation results, combining with the warning index threshold to generate hydrological condition warning information, improving the prediction accuracy and warning timeliness of waves and storm surges in the bay area, and realizing accurate hierarchical response to complex meteorological disasters.

Description

Technical Field

[0001] The present invention relates to the technical field of marine disaster early warning, and particularly to a digital bay hydrological condition early warning method and system. Background Art

[0002] The hydrological conditions in the bay area are affected by various meteorological factors, especially factors such as wind, waves, and storm surges. In order to effectively early warn extreme weather phenomena such as possible storm surges, traditional hydrological prediction methods usually rely on a single meteorological data source and are difficult to comprehensively reflect the interaction between the complex meteorological field and the bay environment. With the progress of computing technology and algorithms, in recent years, composite early warning methods based on multi-source data have gradually received attention.

[0003] This composite early warning method of multi-source data fusion describes the complex bay environment more comprehensively by integrating different meteorological and ocean data sources, especially the interactive influence of the wind field on waves and storm surges. However, the prior art often uses a theoretical typhoon model or a background wind field alone for wind field simulation, resulting in an overestimation of the wind speed in the typhoon core area or the accumulation of errors in the peripheral wind field. For example, Patent CN109583456A proposes to superimpose the theoretical wind field and the background wind field with a fixed weight, but the fixed weight cannot adapt to the dynamic changes of typhoon intensity and distance, resulting in significant errors in the synthesized wind field in the typhoon peripheral area. Moreover, the robustness of the wind-wave relationship model in the prior art is poor. Traditional wave prediction mostly relies on physical models (such as SWAN) or statistical models (such as ARIMA) and is difficult to capture the complex spatio-temporal correlation between wind and waves. For example, Patent CN112950083A uses an LSTM neural network to predict waves, but the model is prone to falling into local optima and has insufficient adaptability to sudden changes in wind direction. Existing machine learning methods do not effectively combine temporal feature extraction (such as wind field gradient) with global parameter optimization, resulting in limited generalization ability of the model in complex sea areas (such as multi-island bays).

[0004] Therefore, there is an urgent need for a new digital bay hydrological condition early warning method and system that can simulate wind field changes at different levels and scales, so as to more accurately predict marine hydrological conditions. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention provides a digital bay hydrological condition early warning method and system, which can improve the prediction accuracy and early warning timeliness of waves and storm surges in the bay area, and achieve accurate hierarchical response to complex meteorological disasters.

[0006] The present invention provides a digital bay hydrological condition early warning method, including the following steps:

[0007] S1. Collect historical meteorological data and real-time meteorological data of the bay area to be predicted;

[0008] The meteorological data includes: wind data, wave data and tidal data;

[0009] S2. According to the historical meteorological data, using deep reinforcement learning and genetic algorithm, establish a relationship model between wind and waves, and generate wave prediction data based on the relationship model and the real-time wind data of the sea area to be predicted;

[0010] S3. Based on the theoretical model wind field and the background wind field, synthesize the synthetic wind field of the sea area to be predicted;

[0011] S4. According to the synthetic wind field and the real-time meteorological data, use the hydrodynamic mathematical model and the tidal current mathematical model to conduct storm surge simulation;

[0012] S5. Based on the wave prediction data and the storm surge simulation results, generate hydrological condition early warning information by combining the early warning index threshold.

[0013] Further, the S2 specifically includes:

[0014] S21. According to the wind data and wave data in the historical meteorological data, pre-train the deep reinforcement learning agent to extract the correlation features between wind and waves;

[0015] S22. Based on the initial individuals recommended by the trained deep reinforcement learning agent and the randomly generated individuals, form an initial population;

[0016] S23. Calculate the fitness of the population individuals according to the wind propagation delay parameter and the wave response time;

[0017] S24. Perform genetic operations on the basis of the fitness through a preset ratio, and perform crossover and mutation operations on the new offspring individuals to generate a new population;

[0018] S25. Repeat steps S22 to S24 until the maximum number of iterations is reached, then stop the iteration and output the relationship model between wind and waves;

[0019] S26. Input the real-time wind data of the sea area to be predicted into the relationship model between wind and waves to generate the wave prediction data of the sea area to be predicted.

[0020] Further, the calculation formula for the fitness in S23 is:

[0021] Take n points in the sea area to be predicted, and divide the influence time into m equal time periods; then the calculation formula for the fitness F is:

[0022] ;

[0023] In the formula, u i,j is the wind speed of the wind data to be predicted; v i,j is the wind direction of the wind data to be predicted; Ui,j is the wind speed of historical wind data; V i,j is the wind direction of historical wind data; i = 1, … n; j = 1, … m; R is the reward value output by the machine learning agent; λ is the adaptive balance coefficient.

[0024] Furthermore, the calculation method of the reward value R is as follows:

[0025] S231. Define the state space as the gradient matrix of the current wind data and the wave data vector [Hs, T];

[0026] S232. The machine learning agent outputs the reward value through the policy network:

[0027] ;

[0028] where w 1, w 2 are network learnable parameters.

[0029] Furthermore, the said S3 specifically includes:

[0030] ;

[0031] In the formula: V c represents the synthetic wind field; V represents the theoretical model wind field; V ERA represents the background wind field; E represents the weight coefficient;

[0032] Furthermore, the calculation of the weight coefficient is as follows:

[0033] ;

[0034] In the formula, d represents the distance from the typhoon center to the target position; d0 represents the reference distance for normalizing the distance d; k represents the attenuation rate coefficient.

[0035] Furthermore, the attenuation rate coefficient k is determined through the following steps: S31. Statistically analyze the maximum influence radius R of historical typhoons in the target sea area max ; S32. Set k = ln2 / (R max / d0)2, and when d = R max E = 0.5.

[0036] Furthermore, the said S4 specifically includes:

[0037] Set the open boundary conditions, closed boundary and moving boundary conditions of the hydrodynamic model according to the tidal forecast data and real-time tidal data;

[0038] According to the synthetic wind field and real-time meteorological data, using the hydrodynamic mathematical model, calculate and output the change of water level over time, the velocity field of the water flow, the surface wind stress, and the bed shear stress;

[0039] Using the data output by the hydrodynamic model, conduct storm surge simulation through the tidal current mathematical model.

[0040] Further, the S5 specifically includes:

[0041] According to the wave prediction data and the storm surge simulation results, determine the early warning indicators of the hydrological conditions and set the early warning thresholds;

[0042] According to the thresholds of the early warning indicators of the hydrological conditions, divide the early warning of the hydrological conditions into 4 early warning levels and formulate corresponding response measures;

[0043] According to the wave prediction data and the storm surge simulation results, determine the early warning level of the bay area to be predicted and implement corresponding measures.

[0044] The present invention also provides a digital bay hydrological condition early warning system for implementing the above-mentioned digital bay hydrological condition early warning method, which is characterized in that the system includes the following modules:

[0045] Data collection module: used to collect historical meteorological data and real-time meteorological data of the bay area to be predicted; the meteorological data includes: wind data, wave data, and tidal data;

[0046] Wave prediction module: connected to the data collection module, used to establish a relationship model between wind and waves according to historical meteorological data by using deep reinforcement learning and genetic algorithm, and generate wave prediction data based on the relationship model and the real-time wind data of the bay area to be predicted;

[0047] Storm surge prediction module: connected to the data collection module, used to synthesize the synthetic wind field of the bay area to be predicted by using the theoretical model wind field and the background wind field; based on the synthetic wind field, combined with real-time meteorological data, conduct storm surge simulation by using the hydrodynamic mathematical model and the tidal current mathematical model;

[0048] Early warning module, connected to the wave prediction module and the storm surge prediction module, generates hydrological condition early warning information based on the wave prediction data and the storm surge simulation results in combination with the early warning index threshold.

[0049] The embodiments of the present invention have the following technical effects:

[0050] 1. By integrating historical meteorological data, real-time meteorological data (such as wind speed, wave data, tidal data) and the output data of the hydrodynamic model, it is possible to more accurately predict the hydrological conditions in the bay area and improve the prediction accuracy of natural disasters such as storm surges and waves.

[0051] 2. By combining the theoretical model wind field and the background wind field, a more accurate synthetic wind field is synthesized. The two are fused with dynamic weights, and the weights decay in a Gaussian type with the distance between the typhoon center and the target position, which is more in line with the actual physical law of typhoon wind speed decay with distance. The attenuation rate coefficient is determined through historical typhoon statistics instead of being preset artificially, enhancing the adaptability of the model to different typhoon intensities. It not only retains the high-precision simulation of the typhoon core but also uses the background field to correct the error of the peripheral wind field, thus providing a more reliable driving wind field for storm surge simulation and reducing the impact of the error of a single wind field model on the simulation results.

[0052] 3. Using deep reinforcement learning (DRL) combined with genetic algorithms, it automatically learns and predicts wave data from a large amount of historical data, greatly shortening the time required for prediction and improving the accuracy of the scheme. DRL dynamically captures the non-linear time-series relationship between wind and waves (such as the lag effect of wind speed mutation on wave height) through a reward mechanism (such as a policy network based on the wind field gradient matrix and wave data), solving the problem of insufficient modeling of time-series dependence in traditional models (such as LSTM, ARIMA); the genetic algorithm screens the optimal parameter combination through a fitness function (fusing wind speed / direction error and DRL reward value), avoiding the model falling into local optimum and enhancing the generalization ability in complex sea areas (such as multi-island areas); through the pre-training of the deep reinforcement learning (DRL) agent, the time-series correlation features of the wind field gradient matrix and wave parameters are extracted to generate the initial population individuals as input to the genetic algorithm, replacing the completely random initialization. Furthermore, it can provide a guiding initial position for the genetic algorithm, which can guide the search process to approach the global optimum solution faster. Compared with the completely random initialization, the initialization based on the predicted values of existing knowledge can guide the search direction and reduce unnecessary blind exploration. Description of the Drawings

[0053] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0054] Figure 1 It is a flowchart of a digital bay hydrological condition early warning method provided by an embodiment of the present invention.

[0055] Figure 2 It is the relative forecast error of wind data in 18 hours of a digital bay hydrological condition early warning method provided by an embodiment of the present invention.

[0056] Figure 3It is the relative prediction error of the wind data in 30 hours of a digital bay hydrographic condition warning method provided by an embodiment of the present invention. Detailed implementation manners

[0057] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below. Apparently, the described embodiments are only some of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope protected by the present invention.

[0058] Figure 1 It is a flowchart of a digital bay hydrographic condition warning method provided by an embodiment of the present invention. Refer to Figure 1 and specifically includes:

[0059] S1. Collect historical meteorological data and real-time meteorological data of the bay area to be predicted;

[0060] The meteorological data includes: wind data, wave data, and tide data.

[0061] Collecting historical meteorological data and real-time meteorological data of the bay area to be predicted is the data cornerstone for constructing a hydrographic warning system. The historical meteorological data covers continuous observation records of the wind field, wave field, and tide field in the past decades, specifically including wind elements such as wind speed and wind direction, wave parameters such as significant wave height, wave period, and propagation direction, and tide information such as astronomical tide level, storm surge process, and tide phase. The real-time meteorological data can be dynamically obtained through meteorological satellite remote sensing, coastal automatic weather stations, buoy arrays, and radar detection equipment.

[0062] S2. According to the historical meteorological data, use deep reinforcement learning and genetic algorithms to establish a relationship model between wind and waves, and generate wave prediction data based on the relationship model and the real-time wind data of the bay area to be predicted;

[0063] In the digital bay hydrographic warning method, establishing a relationship model between wind and waves is the core link, and the key lies in capturing the non-linear correlation between meteorological elements through intelligent algorithms. The implementation process of step S2 integrates the advantages of deep reinforcement learning and genetic algorithms to form a hybrid modeling framework, which not only utilizes the feature extraction ability of reinforcement learning but also gives play to the global optimization characteristics of genetic algorithms, and finally constructs a high-precision wind-wave response model.

[0064] In some embodiments, S2 specifically includes:

[0065] S21. According to the wind data and wave data in the historical meteorological data, pre-train a deep reinforcement learning agent to extract the correlation features between wind and waves.

[0066] First, use the wind data (including wind speed and wind direction) and wave data (such as significant wave height Hs, period T) in historical meteorological data to pre-train the deep reinforcement learning agent. The goal of this stage is to enable the agent to extract the correlation features between wind and waves. In this way, it can capture how wind field changes affect wave parameters and how this impact varies over time and space.

[0067] S22. Form an initial population based on the initial individuals recommended by the trained deep reinforcement learning agent and randomly generated individuals.

[0068] Form an initial population based on the initial individuals recommended by the above-trained deep reinforcement learning agent and randomly generated individuals. This step combines the results of reinforcement learning with the search ability of the genetic algorithm, aiming to provide a diverse starting point for the subsequent optimization process.

[0069] Exemplarily, based on the pre-training results, the agent may recommend a set of parameters as a starting point. For example, it may recommend a set of model parameters that are considered to be able to better simulate the wind-wave relationship based on past experience. Suppose this set of parameters is A = {a1, a2,..., an};

[0070] At the same time, we also randomly generate a set of parameters B = {b1, b2,..., bn}. These parameters represent different model configurations, and they may be randomly selected entirely, aiming to explore different parts of the search space.

[0071] Combine these two sets of parameters, along with a few additional random individuals, to form the initial population. For example, the initial population may contain 5 individuals: A, B, and 3 other randomly generated individuals C, D, and E. The purpose of doing this is to ensure that the initial population has both "experienced" solutions provided by the deep reinforcement learning agent and completely random new attempts, thus ensuring diversity and extensive exploration.

[0072] S23. Calculate the fitness of the population individuals according to the wind propagation delay parameter and the wave response time;

[0073] In some embodiments, the calculation formula for fitness is:

[0074] Take n points in the bay area to be predicted, and divide the influence time into m equal time periods; then the calculation formula for fitness F is:

[0075] ;

[0076] In the formula, u i,j is the wind speed of the wind data to be predicted; v i,j is the wind direction of the wind data to be predicted; U i,j is the wind speed of the historical wind data; Vi,j is the wind direction of historical wind data; i = 1, …, n; j = 1, …, m; R is the reward value output by the machine learning agent; λ is the adaptive balance coefficient.

[0077] The design of the fitness function takes into account multiple points (n) and time intervals (m evenly spaced time periods) within the area to be predicted. The fitness of each individual is calculated by comparing the predicted wind data (u i,j , v i,j ) with the historical wind data (U i,j , V i,j ) and the reward value R output by the machine learning agent, where λ is the adaptive balance coefficient.

[0078] Furthermore, the reward value R is calculated as follows:

[0079] S231. Define the state space as the gradient matrix of the current wind data and the wave data vector [Hs, T];

[0080] S232. The machine learning agent outputs the reward value through the policy network:

[0081] ;

[0082] where, w 1, w 2 are the learnable parameters of the network.

[0083] Optionally, λ, as the adaptive balance coefficient, is used to adjust the relative importance between different factors (such as the influence of historical data and real-time data). It can be determined through cross-validation, grid search / Grid Search, and Bayesian optimization.

[0084] The learnable parameters w 1, w 2 are the learnable parameters in the policy network, which determine how to output the reward value based on the current state (the gradient matrix of wind data and the wave data vector). They can be determined through methods such as backpropagation algorithm, initialization and regularization, and hyperparameter tuning.

[0085] S24. Perform genetic operations on the new offspring individuals according to the fitness by a preset ratio, and generate a new population of offspring through crossover and mutation operations.

[0086] Specifically, the individuals are sorted according to the fitness function, and excellent individuals are selected to enter the next generation according to a preset ratio (e.g., the top 20% of the individuals). This process can adopt methods such as roulette wheel selection and tournament selection. The selected parent individuals are paired, and part of the gene information is exchanged according to a certain crossover probability to generate new offspring. Common crossover methods include single-point crossover, multi-point crossover, and uniform crossover, etc. To increase the population diversity and avoid premature convergence, certain gene values of the newly generated offspring individuals are randomly changed with a certain mutation probability. Mutation can be achieved by replacing a gene value or perturbing it. Through the above genetic operations (selection, crossover, mutation), a new generation of population is generated to prepare for the next iteration.

[0087] S25. Repeat steps S22 - S24 until the maximum number of iterations is reached, then stop the iteration and output the relationship model between wind and waves.

[0088] Specifically, the process of steps S22 to S24 needs to be repeated continuously, that is, recalculate the fitness of each individual in the new generation of population, and then perform selection, crossover, and mutation operations again to generate a new generation of population. Set a maximum number of iterations as the stopping criterion. When the number of iterations reaches the preset maximum value, the algorithm stops running. At this time, the best individual in the output population represents the optimized relationship model between wind and waves.

[0089] S26. Input the real-time wind data of the bay area to be predicted into the relationship model between wind and waves to generate the wave prediction data of the bay area to be predicted.

[0090] Input the real-time wind data (such as wind speed u i,j and wind direction v i,j ) of the bay area to be predicted into the finally obtained relationship model between wind and waves. Based on the input real-time wind data, the model will output the corresponding wave prediction data (such as significant wave height Hs and period T). These prediction data can be used to evaluate the hydrological conditions in the bay area and support the decision-making of the early warning system.

[0091] The above step S2 describes a method of combining deep reinforcement learning and genetic algorithm to establish a relationship model between wind and waves. The core lies in the complementary advantages of the two intelligent algorithms to improve the prediction accuracy. The deep reinforcement learning agent learns the wind speed, wind direction and corresponding wave data (such as significant wave height Hs and period T) in historical meteorological data through the pre-training stage, enabling it to identify how wind field changes affect wave parameters and capture the complex non-linear relationship between the two. On this basis, the genetic algorithm introduces an evolutionary mechanism, evaluates the rationality of different wind propagation delay assumptions through the fitness function, and explores the optimal solution space using crossover and mutation operations. The two cooperate to form a closed loop of "feature extraction - global optimization": reinforcement learning mines local features, and the genetic algorithm realizes parameter tuning, jointly improving the generalization ability of the model.

[0092] Exemplarily, according to the past 30 years of continuous wind data and wave data in the Bohai Bay, the deep reinforcement learning and genetic algorithm are used to compare the current predicted value and the current measured value of a single point. Figure 2 and Figure 3 are the prediction relative errors of wind data at 18 hours and 30 hours respectively.

[0093] S3. Based on the theoretical model wind field and the background wind field, synthesize the synthetic wind field of the bay area to be predicted.

[0094] Step S3 describes how to synthesize the synthetic wind field of the bay area to be predicted based on the theoretical model wind field and the background wind field. This process introduces a weight coefficient E to balance the theoretical wind field (V) and the background wind field (V ERA ), in order to generate a synthetic wind field (V c ) that more accurately reflects the actual conditions. The specific implementation details are as follows:

[0095] In some embodiments, S3 specifically includes:

[0096] ;

[0097] where: V c represents the synthetic wind field; V represents the theoretical model wind field, which represents the wind field predicted based on physical or mathematical models; V ERA represents the background wind field, which usually refers to the climatological mean state or background wind field situation constructed from long-term observation data; E represents the weight coefficient, which is used to adjust the relative contributions between the theoretical model wind field and the background wind field.

[0098] Optionally, the theoretical model wind field is the Holland gradient wind field; the background wind field is the ECMWF background wind field.

[0099] In some embodiments, the calculation of the weight coefficient is as follows:

[0100] ;

[0101] In the formula, d represents the distance from the typhoon center to the target location; d0 represents the reference distance for normalizing the distance d; k represents the attenuation rate coefficient.

[0102] In some embodiments, the attenuation rate coefficient k is determined through the following steps: S31. Statistically analyze the maximum influence radius R of historical typhoons in the target sea area max ; This step involves analyzing typhoon events that occurred in the area in the past, recording the influence range of each event, and calculating an average maximum influence radius as R max . S32. Set k = ln2 / (R max / d0)2, and when d = R max , E = 0.5. The purpose of this step is to ensure that when d = R max , the weight coefficient E = 0.5. This means that within the maximum influence range of the typhoon, the contributions of the theoretical wind field and the background wind field to the synthetic wind field are equal. As the distance increases beyond R max , the influence of the theoretical wind field gradually decreases, while the role of the background wind field increases accordingly.

[0103] Through the above steps, the advantages of the theoretical model wind field and the background wind field can be effectively combined to generate a synthetic wind field model that can not only reflect the dynamic changes of the current weather system but also maintain long-term meteorological characteristics. This method is particularly suitable for complex and changeable marine environments, helps to improve the accuracy of wind wave prediction, and supports more effective hydrological early warning decisions. In addition, by adjusting parameters such as d0 and k, customized adjustments can be made according to the characteristics of different geographical regions, further enhancing the applicability and reliability of the model.

[0104] S4. According to the synthetic wind field and real-time meteorological data, use the hydrodynamic mathematical model and the tidal current mathematical model to conduct storm surge simulations;

[0105] In some embodiments, S4 specifically includes:

[0106] Set the open boundary conditions, closed boundaries, and moving boundary conditions of the hydrodynamic model according to the tidal forecast data and real-time tidal data;

[0107] Open boundary conditions: Set the open boundary conditions of the hydrodynamic model according to tidal prediction data and real-time tidal data. Open boundaries are usually located at the outer boundaries of the study area and are used to describe the interaction between the ocean or large lake and the external water body. These conditions may include parameters such as water level and flow velocity. Closed boundary conditions: Define the boundaries within the study area that are not directly connected to the external water body, such as coastlines, dams, etc. At these boundaries, the water flow velocity is usually set to zero (no-penetration condition), or other physical constraints are set according to specific circumstances. Moving boundary conditions: For some special areas, such as estuaries, wetlands, etc., where there may be periodic wet-dry alternation, moving boundary conditions need to be set to dynamically adjust the boundary position and properties to more accurately simulate the actual environment.

[0108] According to the synthetic wind field and real-time meteorological data, use the hydrodynamic mathematical model to calculate and output the change of water level over time, the velocity field of water flow, the surface wind stress, and the bed shear stress.

[0109] Specifically, take the synthetic wind field and real-time meteorological data as inputs, including but not limited to wind speed, wind direction, air pressure, etc. Use the hydrodynamic mathematical model, based on the above boundary conditions and input data, to calculate and output the change of water level over time, the velocity field of water flow, the surface wind stress, and the bed shear stress. These calculation results are important bases for understanding the impact of storm surges. Water level change: Reflects the change in sea level height during a storm. Water flow velocity field: Displays the water flow direction and intensity at different times and locations. Surface wind stress and bed shear stress: Helps analyze the force of the wind on the water surface and the possibility of seabed sediment transport.

[0110] Use the data output by the hydrodynamic model to conduct storm surge simulation through the tidal current mathematical model.

[0111] Use the data obtained from the hydrodynamic model, especially the water level change and the water flow velocity field, as inputs into the tidal current mathematical model. Further simulate the specific impacts of storm surges through the tidal current mathematical model, including but not limited to the inundation range, flood depth, and its duration in coastal areas, etc. This step helps to evaluate the potential threats of storm surges to human activities and the natural environment and supports the formulation of effective disaster prevention and mitigation strategies.

[0112] S5. Based on the wave prediction data and the storm surge simulation results, generate hydrological condition warning information by combining with the warning index threshold.

[0113] In some embodiments, S5 specifically includes:

[0114] According to the wave prediction data and the storm surge simulation results, determine the hydrological condition warning indicators and set the warning thresholds;

[0115] First, based on wave prediction data (such as significant wave height Hs, period T) and storm surge simulation results (such as maximum inundation depth, flood duration), a series of key hydrological condition warning indicators are determined. These indicators include but are not limited to maximum wave height Hs, average wave period T, highest water level caused by storm surge, flood duration, etc.

[0116] According to the thresholds of the hydrological condition warning indicators, the hydrological condition warning is divided into 4 warning levels and corresponding response measures are formulated;

[0117] Exemplarily, minor warning: when the maximum wave height Hs is between 2 meters and 3 meters; moderate warning: when the maximum wave height Hs is between 3 meters and 4 meters or the highest water level caused by storm surge exceeds the warning water level by less than 0.5 meters; major warning: when the maximum wave height Hs is greater than 4 meters or the highest water level caused by storm surge exceeds the warning water level by more than 0.5 meters; extreme warning: when the maximum wave height Hs exceeds 6 meters or a severe flood disaster is expected to occur.

[0118] Detailed response measures are formulated for each warning level. Exemplarily:

[0119] Minor warning: Strengthen monitoring and notify relevant departments to prepare emergency supplies. Moderate warning: Activate the emergency plan and evacuate residents in low-lying areas. Major warning: Mobilize all rescue forces to ensure the safe evacuation of all personnel. Extreme warning: Urgently close coastal facilities and do everything possible to protect life and property safety.

[0120] Based on the wave prediction data and storm surge simulation results, determine the warning level of the bay area to be predicted and implement corresponding measures.

[0121] Based on the wave prediction data and storm surge simulation results, determine the value of each warning indicator, compare it with the pre-set threshold, determine which warning level the current bay area is in, and immediately execute the corresponding measures at the determined warning level.

[0122] The present invention also provides a digital bay hydrological condition warning system for implementing the above-mentioned digital bay hydrological condition warning method, which is characterized in that the system includes the following modules:

[0123] Data collection module: used to collect historical meteorological data and real-time meteorological data of the bay area to be predicted; the meteorological data includes: wind data, wave data and tidal data;

[0124] Wave prediction module: connected to the data collection module, used to establish a relationship model between wind and waves based on historical meteorological data by using deep reinforcement learning and genetic algorithms, and generate wave prediction data based on the relationship model and the real-time wind data of the bay area to be predicted;

[0125] Storm surge prediction module: Connected to the data collection module, it is used to synthesize the synthetic wind field of the bay area to be predicted by using the theoretical model wind field and the background wind field; Based on the synthetic wind field, combined with real-time meteorological data, the storm surge is simulated by using the hydrodynamic mathematical model and the tidal current mathematical model.

[0126] Early warning module, connected to the wave prediction module and the storm surge prediction module, generates hydrological condition early warning information based on the wave prediction data and the storm surge simulation results, in combination with the early warning index threshold.

[0127] Finally, it should be noted that: The above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting them; Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: They can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A digital early warning method for bay hydrological conditions, characterized in that, The steps include: S1. Collect historical and real-time meteorological data of the bay area to be predicted; The meteorological data includes: wind data, wave data and tidal data; S2. Based on historical meteorological data, deep reinforcement learning and genetic algorithms are used to establish a relationship model between wind and waves, and wave prediction data is generated based on the relationship model and real-time wind data of the bay area to be predicted; The S2 specifically includes: S21. Based on the wind data and wave data in the historical meteorological data, pre-train the deep reinforcement learning agent to extract the correlation features between wind and waves; S22, forming an initial population based on the initial individuals recommended by the trained deep reinforcement learning agent and the randomly generated individuals; S23, calculating the fitness of individuals in the population according to the wind propagation delay parameter and the wave response time; S24, perform genetic operations at a preset ratio according to fitness, and generate offspring populations through crossover and mutation operations on new offspring individuals; S25, repeating steps S22 to S24 until the maximum number of iterations is reached, then stopping the iteration and outputting the relationship model between wind and waves; S26, inputting the real-time wind data of the bay area to be predicted into the relationship model between wind and waves to generate wave prediction data of the bay area to be predicted; The calculation formula of fitness in S23 is: Take n points in the bay area to be predicted and divide the impact time into m uniform time periods; then the calculation formula of fitness F is: ； where u i,j is the wind speed of the wind data to be predicted; v i,j is the wind direction of the wind data to be predicted; U i,j is the wind speed of the historical wind data; V i,j is the wind direction of the historical wind data; i = 1, … n; j = 1, … m; R is the reward value output by the machine learning agent; λ is the adaptive balance coefficient; S3. Based on the theoretical wind field model and the background wind field model, synthesize the synthetic wind field of the bay area to be predicted; The S3 specifically includes: ； Where: V c represents the synthetic wind field; V represents the wind field of the theoretical model; V ERA represents the background wind field; E represents the weight coefficient; S4, performing storm surge simulation using a hydrodynamic mathematical model and a tidal mathematical model according to the synthetic wind field and real-time meteorological data; S5. Generate hydrological condition warning information based on the wave prediction data and storm surge simulation results in combination with warning indicator thresholds.

2. The digital gulf hydrological condition early warning method according to claim 1, wherein, The reward value R is calculated as follows: S231, define the state space as the gradient matrix ∇V of the current wind data and the wave data vector [Hs,T]; S232, the machine learning agent outputs a reward value through a policy network: ； Among them, w 1, w 2 are network learnable parameters.

3. The digital bay hydrographic condition early warning method according to claim 1, characterized in that The weight coefficient is calculated as follows: ； Where d represents the distance from the typhoon center to the target location; d0 represents the reference distance, which is used to standardize the distance d; and k represents the attenuation rate coefficient.

4. A digital bay hydrographic condition early warning method according to claim 3, characterized in that, The attenuation rate coefficient k is determined through the following steps: S31. Statistically analyze the maximum influence radius R of historical typhoons in the target sea area max ; S32. Set k = ln2 / (R max / d0)2, where when d = R max , E = 0.5 5. A method for warning of digital bay hydrological conditions according to claim 1, characterized in that The S4 specifically includes: Set the open boundary conditions, closed boundary conditions and moving boundary conditions of the hydrodynamic model according to the tide forecast data and real-time tide data; Based on the synthetic wind field and real-time meteorological data, the hydrodynamic mathematical model is used to calculate the change of output water level over time, the velocity field of water flow, surface wind stress and bed shear stress; The data output by the hydrodynamic model is used to simulate the storm surge through the tidal mathematical model.

6. The digital gulf hydrographic condition early warning method according to claim 1, characterized in that The S5 specifically includes: Determine hydrological condition warning indicators and set warning thresholds based on wave forecast data and storm surge simulation results; According to the threshold of hydrological condition warning indicators, the hydrological condition warning is divided into 4 warning levels and corresponding response measures are formulated; Based on wave prediction data and storm surge simulation results, the warning level for the bay area to be predicted is determined and corresponding measures are implemented.

7. A digital bay hydrographic condition early warning system for implementing the digital bay hydrographic condition early warning method described in any one of the above claims 1-6, characterized in that, The system includes the following modules: Data collection module: used to collect historical meteorological data and real-time meteorological data of the bay area to be predicted; the meteorological data includes: wind data, wave data and tidal data; Wave prediction module: connected to the data collection module, used to establish a relationship model between wind and waves according to historical meteorological data by using deep reinforcement learning and genetic algorithm, and generate wave prediction data based on the relationship model and real-time wind data of the bay area to be predicted; Storm surge prediction module: connected to the data collection module, used to synthesize the synthetic wind field of the bay area to be predicted by using the theoretical wind field model and the background wind field model; based on the synthetic wind field, combined with real-time meteorological data, use the hydrodynamic mathematical model and the tidal current mathematical model to conduct storm surge simulation; Early warning module, connected to the wave prediction module and the storm surge prediction module, generates hydrological condition early warning information based on the wave prediction data and the storm surge simulation results, in combination with the early warning index threshold.

Images