Sediment section sluice change prediction method under influence of sluice

Abstract

The invention relates to the technical field of sediment section sluicing, in particular to a sediment section sluicing change prediction method under the influence of a water gate. Acquiring a current erosion and deposition topographic sequence corresponding to the target gate at the current moment, and acquiring future rainfall-runoff data, future upstream and downstream water level difference and future tide level change data corresponding to the target gate within a future preset duration; determining future gate scheduling data corresponding to the target gate based on the future rainfall-runoff data, the future upstream and downstream water level difference and the future tide level change data; and inputting the current erosion and deposition topographic sequence, the future rainfall-runoff data, the future upstream and downstream water level difference, the future tide level change data and the future gate scheduling data into a preset erosion and deposition change prediction model, and outputting the silt fracture surface erosion and deposition change corresponding to the target gate in the future preset duration. And timely and reliable technical support is provided for hydraulic scheduling and risk management, and safe and efficient operation of the estuary tide gate is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of sediment cross-section impact, and in particular to a method for predicting changes in sediment cross-section impact under the influence of a sluice. Background Art

[0002] In the operation and management of estuarine tidal gates, accurate prediction of downstream scouring and sedimentation changes is crucial to ensuring the safety and efficient operation of water conservancy projects. The scouring and sedimentation process in the downstream area of ​​estuarine tidal gates is influenced by multiple factors, including gate operation, upstream water flow, downstream tide levels, and sediment sources, and exhibits significant dynamics and complexity. Long-term siltation can cause riverbed uplift, reduce the flood-carrying cross-section of the river channel, severely inhibit the effective flood discharge capacity of the gates, and easily cause flood disasters during flood season. Long-term scouring can expose or damage the bottom structure of the gates, creating risks similar to bridge pier scouring, threatening the structural safety of the tidal gates and, in turn, affecting the proper functioning of regional flood control, drainage, and water resource utilization.

[0003] At present, traditional methods for analyzing scouring and sedimentation changes in the downstream area of ​​estuary tidal gates mainly rely on numerical models such as Delft3D. This type of model is based on hydrodynamics and sediment transport theory, and can achieve high prediction accuracy by accurately depicting the water flow movement and sediment exchange process. However, its operation requires a large amount of computing resources and a long computing time, and the model parameter calibration and verification process is cumbersome, resulting in high operating costs. In actual water conservancy project management, there is often a need for rapid evaluation of scheduling strategies and rolling forecasts of multiple schemes. For example, when responding to sudden floods and formulating water resource allocation plans, it is necessary to output the prediction results of scouring and sedimentation changes under multiple scheduling schemes in a short period of time to assist decision-making. However, traditional numerical models are difficult to meet the timeliness requirements in such scenarios.

[0004] Therefore, there is an urgent need for an efficient method to predict the changes in sediment cross-section under the influence of sluice gates that can adapt to multi-source uncertain inputs, so as to provide timely and reliable technical support for hydraulic engineering scheduling and risk management, and realize the safe and efficient operation of estuary tidal gates. Summary of the Invention

[0005] In view of this, the present invention provides a method for predicting changes in sediment cross-section impact under the influence of a sluice gate, so as to solve the problem of the urgent need for a method for predicting changes in sediment cross-section impact under the influence of a sluice gate that is efficient and adaptable to multi-source uncertain inputs.

[0006] In a first aspect, the present invention provides a method for predicting changes in sediment cross-section impact under the influence of a sluice gate, the method comprising:

[0007] Obtain the current scouring and silting terrain sequence corresponding to the target gate at the current moment, and obtain the future precipitation-runoff data, future upstream and downstream water level difference, and future tidal level change data corresponding to the target gate within a preset time period in the future;

[0008] Determine the future gate operation data corresponding to the target gate based on future precipitation-runoff data, future upstream and downstream water level difference and future tidal level change data;

[0009] The current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into the preset scouring and deposition change prediction model, and the sediment section scouring and deposition changes within the future preset time period corresponding to the target gate are output.

[0010] The embodiment of the present application provides a method for predicting changes in sediment cross-section scouring and deposition under the influence of a sluice gate, which obtains the current scouring and deposition topography sequence corresponding to the target gate at the current moment, achieves precise focus on a specific area, avoids redundant interference from global data, and ensures that the model input data is strongly correlated with the target scenario. It obtains precipitation-runoff data, future upstream and downstream water level differences, and future tidal level change data within a preset future time period, comprehensively covering the core hydrological elements that affect sediment scouring and deposition. These data represent potential influencing factors in the future, enabling the model to deduce scouring and deposition changes based on forward-looking information, improve the lead time and comprehensiveness of predictions, and meet the needs of water conservancy decision-makers for predicting future trends.

[0011] Then, based on future precipitation-runoff data, future upstream and downstream water level differences, and future tidal level fluctuations, future gate scheduling data corresponding to the target gate is determined, changing the traditional experience-based scheduling model. By quantitatively analyzing the impact of various factors on gate scheduling, dynamic and scientific scheduling decisions can be made. For example, when heavy rainfall is predicted to cause a high water level difference, gate opening adjustments can be planned in advance, improving flood control capabilities and water resource allocation efficiency.

[0012] Next, the pre-set erosion and deposition prediction model is fed with the current erosion and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level differences, future tidal level changes, and future gate operation data. This integrates historical topography, future environmental variables, and scheduling decision-making factors to form a multi-dimensional input system. These different data complement and validate each other: for example, current topography determines the initial sediment state, future precipitation-runoff data drives the evolutionary process, and gate operation regulates water flow and sediment transport. Together, these data enhance the model's ability to simulate complex erosion and deposition processes. The pre-set erosion and deposition prediction model outputs erosion and deposition changes along a predetermined timeframe, providing intuitive and quantitative forecasts for water conservancy project management. Based on erosion and deposition trends, managers can plan engineering measures such as river dredging and embankment reinforcement, or optimize water resource allocation plans. Furthermore, the forecast results can be used to validate scheduling decisions, forming a virtuous cycle of "data-decision-prediction-feedback," enhancing the scientific and reliable operation of water conservancy systems. This provides timely and reliable technical support for hydraulic engineering scheduling and risk management, ensuring the safe and efficient operation of estuary tidal gates.

[0013] In an optional embodiment, the preset scour change prediction model includes an scour and deposition trend classification model and a sediment scour change prediction model; the current scour and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into the preset scour change prediction model, and the output of the sediment cross-section scour change corresponding to the target gate within a preset time period in the future includes:

[0014] The current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data are input into the scouring and deposition trend classification model to determine the sediment evolution trend category corresponding to the target gate;

[0015] The sediment evolution trend category, current scouring and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into the preset sediment scouring change prediction model, and the sediment scouring change of the target gate within the preset time period in the future is output.

[0016] In an optional embodiment, the current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data are input into the scouring and deposition trend classification model to determine the sediment evolution trend category corresponding to the target gate, including:

[0017] Extract features of the current scouring and silting terrain sequence to obtain terrain data features corresponding to the current scouring and silting terrain sequence;

[0018] Fusion of future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate operation data to generate target fusion features;

[0019] The terrain data features and target fusion features are input into the scouring and deposition trend classification model to determine the sediment evolution trend category corresponding to the target gate.

[0020] In an optional embodiment, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are fused to generate target fusion features, including:

[0021] Based on a preset adaptive phase space reconstruction algorithm, the chaotic characteristics corresponding to future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate operation data are analyzed. The chaotic characteristics include the maximum Lyapunov exponent, fractal dimension, and chaotic frequency spectrum.

[0022] The chaotic features corresponding to future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate operation data are input into the reinforcement learning model in the scouring and deposition trend classification model to determine the target fusion strategy corresponding to each chaotic feature.

[0023] Based on the target fusion strategy, each chaotic feature is fused to generate the target fusion feature.

[0024] In an optional embodiment, the terrain data features and the target fusion features are input into the scouring and sedimentation trend classification model to determine the sediment evolution trend category corresponding to the target gate, including:

[0025] Input terrain data features and target fusion features into the quantum decision tree in the erosion and deposition trend classification model; each decision node in the quantum decision tree is an independent quantum system;

[0026] Initialize the quantum state of each decision node in the quantum decision tree;

[0027] The input terrain data features and target fusion features are encoded using quantum bits to generate the current quantum state. Each sub-feature in the target fusion feature corresponds to one or more quantum bits. A quantum bit can represent multiple states simultaneously.

[0028] Perform quantum gate operations on the current quantum state to generate node quantum states corresponding to each decision node;

[0029] Calculate the quantum probability amplitude of the node quantum state corresponding to each decision node;

[0030] Determine the classification path based on each quantum probability amplitude;

[0031] According to each classification path, enter the next layer of decision nodes until reaching the leaf node;

[0032] The leaf node determines the sediment evolution trend category corresponding to the target gate based on the previously calculated quantum probability amplitude and path information.

[0033] In an optional implementation, performing a quantum gate operation on the current quantum state to generate a node quantum state corresponding to each decision node includes:

[0034] Identify the current quantum state, determine the real-time dynamic features corresponding to the current quantum state, as well as the core features and auxiliary features of the current quantum state;

[0035] Based on real-time dynamic characteristics, determine the target quantum gate combination corresponding to the current quantum state;

[0036] Perform entanglement calculation on the core features based on the controlled NOT gate in the target quantum gate combination to generate high-order entanglement features;

[0037] Based on the controlled NOT gate in the target quantum gate combination, the auxiliary features are entangled with the core features to form a star topology structure;

[0038] Perform nonlinear transformation on the current quantum state to generate nonlinear characteristics corresponding to the current quantum state;

[0039] Based on high-order entanglement characteristics, star topology and nonlinear characteristics, the node quantum state corresponding to each decision node is generated.

[0040] In an optional embodiment, the sediment evolution trend category, the current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into a preset sediment scouring change prediction model, and the sediment scouring change of the target gate within a preset time period in the future is output, including:

[0041] Input sediment evolution trend categories, current scouring and deposition topography sequences, future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate operation data into the time-dependent network of the pre-set sediment scouring change prediction model, and output the time-dependent features.

[0042] Input sediment evolution trend categories, current erosion and deposition topography sequences, future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate operation data into the structural feature extraction network in the pre-set sediment erosion change prediction model to output topological structural features;

[0043] Fuse the time-dependent features and topological structure features and output the fused prediction features;

[0044] Based on the integrated prediction features, the sediment cross-section changes corresponding to the target gate within the preset time period in the future are output.

[0045] In an optional embodiment, the sediment evolution trend category, the current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data are input into a time-dependent network in a preset sediment scouring and change prediction model, and the output time-dependent features include:

[0046] Input sediment evolution trend categories, current erosion and deposition topography sequences, future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate operation data into the time-dependent network of the pre-set sediment erosion change prediction model;

[0047] The weight determination network in the time-dependent network evaluates the input data at each time step to determine the importance of the input data at each time step to the current erosion and deposition topography sequence and the changes in sediment cross-section erosion within a preset time period in the future;

[0048] Determine the weight information corresponding to the input data at each time step according to the importance of the input data at each time step;

[0049] Dynamically adjust the update gate and forget gate in the time-dependent network based on the weight information corresponding to the input data at each time step;

[0050] According to the update gate and forget gate adjusted at each time step, time series features are extracted from the input data and time-dependent features are output.

[0051] In an optional embodiment, the sediment evolution trend category, the current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data are input into the structural feature extraction network of the preset sediment scouring change prediction model to output topological structural features, including:

[0052] The target graph structure is constructed based on the current erosion and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data.

[0053] The graph structure is extracted based on the structural feature extraction network, and the topological structure features are output.

[0054] In an optional embodiment, based on future precipitation-runoff data, future upstream and downstream water level difference and future tidal level change data, determining future gate scheduling data corresponding to the target gate includes:

[0055] Obtain historical precipitation-runoff data, historical upstream and downstream water level difference, historical tidal level change data, and historical gate dispatching data corresponding to historical precipitation-runoff data, historical upstream and downstream water level difference, and historical tidal level change data;

[0056] Based on historical precipitation-runoff data, historical upstream and downstream water level difference, historical tidal level change data, and historical gate operation data, a causal graph structure is constructed;

[0057] The causal graph structure is extended to a dynamic structural equation model to capture the time-lagged causal relationship between variables;

[0058] Substitute future precipitation-runoff data, future upstream and downstream water level differences, and future tidal level change data into the time-lag causal relationship to determine the future gate scheduling data corresponding to the target gate. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0060] Figure 1 2 is a flow chart of a method for predicting changes in sediment cross-section flow under the influence of a sluice gate according to an embodiment of the present invention;

[0061] Figure 2 FIG. 2 is a flow chart of another method for predicting changes in sediment cross-section impact under the influence of a sluice gate according to an embodiment of the present invention. DETAILED DESCRIPTION

[0062] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0063] It should be noted that the method for predicting changes in sediment cross-section provided in the embodiment of the present application may be executed by a device for predicting changes in sediment cross-section, and the device for predicting changes in sediment cross-section may be implemented as part or all of a computer device through software, hardware, or a combination of software and hardware, wherein the computer device may be a server or a terminal, wherein the server in the embodiment of the present application may be a single server or a server cluster composed of multiple servers, and the terminal in the embodiment of the present application may be a smart phone, personal computer, tablet computer, wearable device, smart robot or other smart hardware device. In the following method embodiments, the execution subject is an electronic device as an example for explanation.

[0064] According to an embodiment of the present invention, an embodiment of a method for predicting changes in sediment cross-section impact under the influence of a sluice gate is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0065] In this embodiment, a method for predicting the change of sediment cross-section under the influence of a sluice gate is provided, which can be used in the above-mentioned electronic device. Figure 1FIG. 1 is a flow chart of a method for predicting changes in sediment cross-section under the influence of a sluice gate according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0066] Step S101, obtaining the current scouring and silting terrain sequence corresponding to the target gate at the current moment, and obtaining the future precipitation-runoff data, future upstream and downstream water level difference and future tidal level change data corresponding to the target gate within a future preset time period.

[0067] Specifically, the electronic device can receive the current scouring and silting terrain sequence corresponding to the target gate at the current moment input by the user, and obtain the future precipitation-runoff data, future upstream and downstream water level difference and future tide level change data within the future preset time period corresponding to the target gate. It can also receive the current scouring and silting terrain sequence corresponding to the target gate at the current moment sent by other devices, and obtain the future precipitation-runoff data, future upstream and downstream water level difference and future tide level change data within the future preset time period corresponding to the target gate.

[0068] Optionally, the electronic device can also collect the terrain point cloud data corresponding to the target gate at the current moment based on sonar sounding technology. The terrain point cloud data is processed to generate a continuous and smooth current scouring and silting terrain sequence. The electronic device can obtain the future precipitation data corresponding to the target gate within the future preset time period based on the weather preset website, and obtain the future runoff number corresponding to the target gate within the future preset time period based on the hydrological monitoring station, and generate the future precipitation-runoff data corresponding to the target gate within the future preset time period. The electronic device can also predict the future upstream and downstream water level difference corresponding to the target gate within the future preset time period based on the water level monitoring system, and predict the future tide level change data corresponding to the target gate within the future preset time period based on the tide level monitoring station.

[0069] The embodiment of the present application does not specifically limit the manner in which the electronic device obtains the current scouring and deposition terrain sequence corresponding to the target gate at the current moment, and obtains the future precipitation-runoff data, future upstream and downstream water level difference and future tide level change data within a preset future time period corresponding to the target gate.

[0070] Step S102: Determine the future gate dispatching data corresponding to the target gate based on the future precipitation-runoff data, the future upstream and downstream water level difference and the future tidal level change data.

[0071] Specifically, the electronic device can determine the future gate scheduling data corresponding to the target gate based on the correlation between future precipitation-runoff data, future upstream and downstream water level differences, future tide level change data and future gate scheduling data.

[0072] This step will be described in detail below.

[0073] In step S103, the current scouring and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into a preset scouring and deposition change prediction model, and the scouring and deposition change of the sediment section corresponding to the target gate within the future preset time period is output.

[0074] Specifically, the electronic device can input the current scouring and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tide level change data and future gate scheduling data into a preset scouring and deposition change prediction model. The preset scouring and deposition change prediction model extracts features from the input data and, based on the extracted features, outputs the scouring and deposition changes of the sediment section corresponding to the target gate within a preset time period in the future.

[0075] This step will be described in detail below.

[0076] The embodiment of the present application provides a method for predicting changes in sediment cross-section scouring and deposition under the influence of a sluice gate, which obtains the current scouring and deposition topography sequence corresponding to the target gate at the current moment, achieves precise focus on a specific area, avoids redundant interference from global data, and ensures that the model input data is strongly correlated with the target scenario. It obtains data on precipitation-runoff, future upstream and downstream water level differences, and future tidal level changes within a preset time period in the future, comprehensively covering the core hydrological elements that affect sediment scouring and deposition. These data represent potential influencing factors in the future, enabling the model to deduce scouring and deposition changes based on forward-looking information, improve the lead time and comprehensiveness of predictions, and meet the needs of water conservancy decision-makers for predicting future trends.

[0077] Then, based on future precipitation-runoff data, future upstream and downstream water level differences, and future tidal level fluctuations, future gate scheduling data corresponding to the target gate is determined, changing the traditional experience-based scheduling model. By quantitatively analyzing the impact of various factors on gate scheduling, dynamic and scientific scheduling decisions can be made. For example, when heavy rainfall is predicted to cause a high water level difference, gate opening adjustments can be planned in advance, improving flood control capabilities and water resource allocation efficiency.

[0078] Next, the pre-set erosion and deposition prediction model is fed with the current erosion and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level differences, future tidal level fluctuations, and future gate operation data. This multi-dimensional input system integrates historical topography, future environmental variables, and scheduling decision factors. These different data sets complement and validate each other. For example, current topography determines the initial sediment state, future precipitation-runoff data drives the evolutionary process, and gate operation regulates water flow and sediment transport. Together, these data enhance the model's ability to simulate complex erosion and deposition processes. The pre-set erosion and deposition prediction model outputs erosion and deposition changes at a specific cross-section over a pre-set timeframe, providing intuitive and quantitative forecasts for water conservancy project management. Based on erosion and deposition trends, managers can plan engineering measures such as river dredging and embankment reinforcement, or optimize water resource allocation plans. Furthermore, the forecast results can be used to validate scheduling decisions, forming a virtuous cycle of "data-decision-prediction-feedback," enhancing the scientific and reliable operation of water conservancy systems.

[0079] In this embodiment, a method for predicting the change of sediment cross-section under the influence of a sluice gate is provided, which can be used in the above-mentioned electronic device. Figure 2 FIG. 1 is a flow chart of a method for predicting changes in sediment cross-section under the influence of a sluice gate according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0080] Step S201, obtain the current scouring and deposition terrain sequence corresponding to the target gate at the current moment, and obtain future precipitation-runoff data, future upstream and downstream water level difference and future tidal level change data within a future preset time period corresponding to the target gate.

[0081] For details about this step, please refer to the above description of step S201 and will not be repeated here.

[0082] Step S202: Determine the future gate dispatching data corresponding to the target gate based on the future precipitation-runoff data, the future upstream and downstream water level difference and the future tidal level change data.

[0083] Specifically, the above step S202 may include the following steps:

[0084] Step S2021, obtaining historical precipitation-runoff data, historical upstream and downstream water level difference, historical tidal level change data, and historical gate scheduling data corresponding to the historical precipitation-runoff data, historical upstream and downstream water level difference, and historical tidal level change data.

[0085] Specifically, the electronic device can receive historical precipitation-runoff data, historical upstream and downstream water level differences, historical tidal change data, and historical gate scheduling data corresponding to historical precipitation-runoff data, historical upstream and downstream water level differences, and historical tidal change data input by the user. It can also receive historical precipitation-runoff data, historical upstream and downstream water level differences, historical tidal change data, and historical gate scheduling data corresponding to historical precipitation-runoff data, historical upstream and downstream water level differences, and historical tidal change data sent by other devices. It can also search for historical precipitation-runoff data, historical upstream and downstream water level differences, historical tidal change data, and historical gate scheduling data corresponding to historical precipitation-runoff data, historical upstream and downstream water level differences, and historical tidal change data from the storage space.

[0086] The embodiments of the present application do not specifically limit the manner in which electronic devices obtain historical precipitation-runoff data, historical upstream and downstream water level differences, historical tidal change data, and historical gate scheduling data corresponding to historical precipitation-runoff data, historical upstream and downstream water level differences, and historical tidal change data.

[0087] Step S2022: construct a causal graph structure based on historical precipitation-runoff data, historical upstream and downstream water level differences, historical tide level change data, and historical gate scheduling data.

[0088] Specifically, the electronic device uses a constraint-based method to construct a preliminary causal network skeleton by using the statistical relationship between variables in historical data through conditional independence tests (such as chi-square test and Fisher'sz test). For example, first determine whether precipitation and runoff are unconditionally independent. If not, then test whether they are independent under given other variables (such as soil moisture), gradually eliminate false associations, and determine potential causal edges. During the identification process, the algorithm discovers key thresholds through data statistical analysis. For example, when the distribution of historical data on precipitation intensity shows obvious segmented characteristics, the algorithm can automatically divide the threshold interval and find that when the precipitation intensity exceeds a certain threshold (such as 50mm / h), the influence coefficient of runoff on the water level difference increases significantly. This conditional causal relationship is added to the causal graph to improve the graph structure.

[0089] In addition, a quarterly evaluation cycle is established, with major hydrological events such as typhoons and heavy rains serving as triggers. After an event occurs, a data collection process is immediately initiated, integrating various hydrological data from before and after the event into a historical database. At the evaluation point, the causal discovery algorithm is rerun, combining the newly added data with historical data for analysis. The current causal graph structure is compared with the reanalysis results to detect new causal paths. For example, after a heavy rainstorm, a sudden increase in river sediment concentration may be observed, forming a new causal chain with precipitation intensity and runoff velocity. A reward function is then constructed using the actual gate operation results as feedback. If the operation plan guided by the current causal graph effectively reduces flood risk, the weight of the relevant causal edge is increased; if it causes water resource waste or ecological problems, the corresponding edge weight is decreased. Through continuous iteration using reinforcement learning algorithms (such as Q-learning), the weights of the causal graph edges are more closely aligned with actual physical processes. For example, the weights of the causal edge between tide level and gate operation can be dynamically adjusted to adapt to the varying impacts of seasonal tidal changes on operation, thereby generating the final causal graph structure.

[0090] Step S2023 : Expand the causal graph structure into a dynamic structural equation model to capture the time-lagged causal relationship between variables.

[0091] Specifically, the nodes in the causal graph structure (such as precipitation, runoff, water level difference, and gate scheduling) are converted into variables in the dynamic structural equation model, and the directed edges are converted into causal paths between variables. Each variable is identified to determine the endogenous variables (such as gate scheduling data and water level difference) and exogenous variables (such as precipitation and tidal changes).

[0092] For each endogenous variable, an autoregressive (AR) term and a moving average (MA) term are introduced. For example, for the target gate scheduling data yt, an AR(p) model is constructed: in is the autoregressive coefficient, ∈ t is the random error term; at the same time, the MA(q) term is introduced to consider the impact of past errors on the current value, that is, Capture time series dependencies of variables.

[0093] Then, the spatial correlation strength between each monitoring point is used to define the spatial weight matrix W. For example, the weight is determined based on the inverse of the Euclidean distance, and the closer the monitoring point is, the higher the weight is. The spatial lag term is introduced into the model. For the upstream water level variable x s,t (s represents spatial position, t represents time), considering its spatial lag effect: Σ s W ss′ x s′，t , represents the weighted sum of the water levels at all monitoring points upstream of the target gate. The time lag term is combined with the spatial lag term to form a spatiotemporal lag term, which is included in the structural equation model.

[0094] Integrating the above time delay and spatial location factors into the traditional structural equation model, we obtain an extended dynamic structural equation model. For example, the prediction equation for the target gate scheduling data yt can be expressed as:

[0095]

[0096] Among them, γ s and δ s,i is the spatial position related parameter, ω s is the spatial lag coefficient, which comprehensively describes the spatiotemporal causal relationship between variables.

[0097] Step S2024: Substitute future precipitation-runoff data, future upstream and downstream water level differences, and future tidal level change data into the time-lag causal relationship to determine future gate scheduling data corresponding to the target gate.

[0098] Specifically, the electronic device can substitute future precipitation-runoff data, future upstream and downstream water level differences, and future tide level change data into the time-delay causal relationship based on the set multi-objective function and constraints to determine the future gate scheduling data corresponding to the target gate.

[0099] Specifically, with flood risk reduction as the core, the goal is quantified into indicators such as minimizing the flooded area and minimizing the duration of exceeding the warning water level. For example, f1 is used to represent the flooded area, and the flooding range under different gate scheduling schemes is simulated through a hydraulic model and incorporated into the objective function. For example, In this formula, the first half sums the flooded area at each time step, and the second half determines whether the actual water level exceeds the warning level and performs a weighted summation of the duration of the exceedance to comprehensively reflect the degree of flood risk. To maximize the effective use of water resources, indicators include agricultural irrigation water satisfaction rate, industrial water supply, etc. For example, f2 represents the total water supply in the region, and is calculated by counting the actual water withdrawal of each water-using department under different scheduling schemes. For example, This formula calculates the sum of the ratios of the actual agricultural and industrial water supply to the target supply at each time step. The larger the ratio, the higher the water resource utilization efficiency. Focus on maintaining the ecological flow of the river and protecting biological habitats, with the goal of minimizing the degree to which the ecological flow deviates from the ideal value. For example, set the ecological base flow threshold of the river, and use f3 to represent the total deviation of the actual flow from the ecological base flow threshold. For example,

[0100] Among them, the constraints include: Physical constraints: Consider the physical limitations of the gate itself, such as the gate opening range (0≤ui≤umax, where ui is the gate opening at the i-th time step), the upper and lower water level limits, etc. Water balance constraints: Based on the principle of water conservation within the basin, an equation is established to ensure the balance of water inflow and outflow within each time period, such as It-Ot=ΔSt, where It is the inflow, Ot is the outflow, and ΔSt is the change in water storage within the time period. Water quality constraints: To meet the needs of water pollution control, the concentration of discharged pollutants is limited to not exceeding environmental standards, such as Cpollutant≤Climit.

[0101] Each particle is then considered a potential gate scheduling solution. The particle's position vector corresponds to the gate opening parameter at different time steps, and the velocity vector determines the position update direction and step size. During initialization, the particle's position and velocity are randomly distributed within the feasible solution space.

[0102] Convert the multi-objective function into a fitness function, integrating multiple objectives through weighted summation, such as Fitness = w1f1 + w2f2 + w3f3, where w1, w2, and w3 are the weights of each objective, and w1 + w2 + w3 = 1. Weights can be determined using the Analytic Hierarchy Process or expert scoring.

[0103] For solutions that do not meet the constraints, a penalty term is set to reduce their fitness. For example, a certain penalty score is subtracted from the fitness value for each violation of a constraint.

[0104] The particle speed and position are updated based on the individual optimal position (pbest) and the global optimal position (gbest). Each particle learns from its own historical optimal solution and the global optimal solution, continuously adjusting its position in the solution space, gradually approaching the optimal scheduling solution, and obtaining the future gate scheduling data corresponding to the target gate.

[0105] In step S203, the current scouring and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into a preset scouring and deposition change prediction model, and the scouring and deposition change of the sediment section corresponding to the target gate within the future preset time period is output.

[0106] Specifically, the preset erosion change prediction model includes an erosion and deposition trend classification model and a sediment erosion change prediction model; the above step S203 may include the following steps:

[0107] In step S2031, the current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into the scouring and deposition trend classification model to determine the sediment evolution trend category corresponding to the target gate.

[0108] Specifically, the above step S2031 may include the following steps:

[0109] Step a1: extract features of the current scouring and silting terrain sequence to obtain terrain data features corresponding to the current scouring and silting terrain sequence.

[0110] Specifically, the preset erosion change prediction model can calculate the average elevation, elevation standard deviation, maximum elevation, minimum elevation, etc. corresponding to the current erosion and deposition terrain sequence. Then, the slope and slope direction are calculated by differentiating the average elevation, elevation standard deviation, maximum elevation, and minimum elevation. The preset erosion change prediction model can also calculate the curvature characteristics (including plan curvature and profile curvature), the roughness of the terrain surface, and the fractal dimension corresponding to the current erosion and deposition terrain sequence. Among them, the roughness can be calculated based on the rate of change of elevation data, such as using the sum of the absolute values ​​of the elevation differences between adjacent points. Fractal dimension: Fractal dimension can be used to measure the complexity of the terrain. It reflects the self-similarity of the terrain at different scales, and is usually calculated using methods such as the box dimension. The larger the fractal dimension, the more complex the terrain.

[0111] Furthermore, the pre-set erosion and deposition change prediction model can calculate the change in erosion and deposition terrain between adjacent time steps, yielding the erosion and deposition rate. By analyzing the magnitude and distribution of erosion and deposition rates, we can understand the speed and trend of terrain change and identify areas of active erosion and deposition and stable areas. By calculating the area of ​​regions with varying degrees of erosion and deposition, for example, by dividing erosion and deposition terrain into accumulation zones, scour zones, and stable zones, we can calculate the area of ​​each zone and its proportion of the total area to understand the spatial distribution of erosion and deposition.

[0112] Finally, the characteristic values ​​such as elevation statistical characteristics, slope and aspect, curvature characteristics, roughness, fractal dimension, scouring and silting rate, and scouring and silting area statistics are arranged in a certain order to form a multidimensional feature vector as the terrain data feature corresponding to the current scouring and silting terrain sequence.

[0113] In step a2, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data are integrated to generate target fusion features.

[0114] Specifically, the above step a2 may include the following steps:

[0115] Step a21, based on a preset adaptive phase space reconstruction algorithm, analyzes the chaotic characteristics corresponding to the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data.

[0116] Among them, the chaotic characteristics include the maximum Lyapunov exponent, fractal dimension and chaotic frequency spectrum.

[0117] Specifically, for future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate scheduling data, the phase space is constructed using a preset adaptive phase space reconstruction algorithm based on Takens' theorem.

[0118] Specifically, we select an appropriate embedding dimension m and time delay τ to map the time series data into phase space. For example, for a precipitation-runoff data series {x(t)}, the reconstructed phase space vector is X(t) = [x(t), x(t+τ), …, x(t+(m-1)τ)]. m and τ are dynamically adjusted through an adaptive algorithm to adapt to the characteristics and variations of the data.

[0119] In the reconstructed phase space, calculate the maximum Lyapunov exponent (λmax). The maximum Lyapunov exponent measures the rate of separation between adjacent orbits in the phase space, reflecting the system's degree of chaos. Alternatively, the Wolf algorithm can be used to track the evolution of orbits in phase space and calculate the rate of growth of the distance between them, thereby obtaining the maximum Lyapunov exponent. When λmax > 0, the system exhibits chaotic characteristics; larger λmax indicates a higher degree of chaos.

[0120] Fractal dimension is used to describe the complexity of attractors in phase space. Specifically, the box dimension calculation method can be used to divide the phase space into boxes of different scales. The number of boxes N(∈) containing attractors can be counted. By analyzing N(∈) at different scales ∈, the fractal dimension D is obtained. The larger the fractal dimension D, the more complex the attractor structure and the more pronounced the chaotic characteristics of the system.

[0121] Finally, spectral analysis is performed on future precipitation-runoff data, future upstream-downstream water level differences, future tidal changes, and future gate operation data. For example, a fast Fourier transform (FFT) is used to convert time series data into the frequency domain to obtain a chaotic frequency spectrum. The chaotic frequency spectrum shows the energy distribution of the system at different frequencies. By analyzing the characteristics of the frequency spectrum, such as peak location and bandwidth, we can further understand the chaotic behavior and dynamic characteristics of the system.

[0122] Chaotic characteristics are generated based on the maximum Lyapunov exponent, fractal dimension and chaotic frequency spectrum.

[0123] In step a22, the chaotic features corresponding to the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data are input into the reinforcement learning model in the scouring and deposition trend classification model to determine the target fusion strategy corresponding to each chaotic feature.

[0124] Specifically, the electronic device inputs chaotic features corresponding to future precipitation-runoff data, future upstream-downstream water level differences, future tidal changes, and future gate operation data into the reinforcement learning model within the erosion and deposition trend classification model. The reinforcement learning model constructs a state space based on the input data and an action space. This action space encompasses various possible chaotic feature fusion methods. Taking weighted fusion as an example, an action can be defined as an operation that assigns weights to different chaotic features. For example, action 1 = [0.3, 0.4, 0.3] represents weights of 0.3, 0.4, and 0.3 assigned to the maximum Lyapunov exponent, fractal dimension, and chaotic frequency spectrum, respectively. This also includes the selection and parameter settings of other fusion algorithms, such as principal component analysis (PCA) fusion and independent component analysis (ICA) fusion. Each time, the reinforcement learning model selects an action from the action space and calculates the corresponding reward function. The reward function can focus on improving the accuracy of erosion and deposition trend predictions while also considering the effectiveness and stability of the fused features. For example, if the features after model fusion are used in the erosion and siltation trend prediction model, when the error between the prediction result and the actual erosion and siltation situation is small, a positive reward is given; on the contrary, if the prediction error is large, a negative reward is given. For example, multiple indicators can be used to quantify the reward. For example, based on the mean square error (MSE) calculation, assuming that the predicted erosion and siltation amount is The actual erosion and deposition volume is yi, and there are n data points in a prediction. The reward value R can be set as R = 1 / 1 + MSE, so that the smaller the MSE, the larger the reward. In addition, other indicators can be combined, such as the accuracy of the predicted trend direction (a reward is given if the predicted sedimentation or scour direction is consistent with the actual direction) and the variance of the fused features (a small variance indicates high feature stability and is rewarded).

[0125] Then, according to the reward system corresponding to each action, each action is updated based on the policy function π(a|s). Among them, the policy function π(a|s) represents the probability of taking action a in state s. Common policy representations include parameter-based policy networks, such as using neural networks to construct policy functions, with the input being the state (chaotic characteristics) and the output being the probability distribution of each action. The parameters of the policy network are adjusted based on reward feedback through gradient ascent algorithms (such as policy gradient algorithms), so that the probability of actions that can obtain high rewards being selected increases. For example, in a certain state, if a fusion action with a specific weight distribution obtains a high reward, the policy network will adjust the parameters to increase the probability of the action being selected in similar states.

[0126] Finally, the action with the highest reward value is determined as the target fusion strategy corresponding to each chaotic feature.

[0127] Step a23: fuse the chaotic features based on the target fusion strategy to generate the target fusion feature.

[0128] Specifically, the chaotic features are fused based on the target fusion strategy to generate the target fusion feature. For example, if the target fusion strategy is weighted fusion, the maximum Lyapunov exponent, fractal dimension, chaotic frequency spectrum and other features are weighted and summed according to the weights of different chaotic features to obtain the fused feature vector.

[0129] Step a3: Input the terrain data features and target fusion features into the scouring and deposition trend classification model to determine the sediment evolution trend category corresponding to the target gate.

[0130] Specifically, the above step a3 may include the following steps:

[0131] Step a31: input the terrain data features and target fusion features into the quantum decision tree in the erosion and deposition trend classification model.

[0132] Among them, each decision node in the quantum decision tree is an independent quantum system.

[0133] Specifically, the terrain data features and the generated target fusion features are input into the quantum decision tree in the erosion and siltation trend classification model. Since each decision node in the quantum decision tree is an independent quantum system, these features serve as input data and provide the quantum decision tree with basic information for analyzing erosion and siltation trends. For example, terrain data features may include information such as slope and elevation, and target fusion features may incorporate multiple aspects of information such as chaotic features. Together, they provide a basis for the decision-making process of the decision tree.

[0134] Step a32: Initialize the quantum state of each decision node in the quantum decision tree.

[0135] Specifically, the quantum state of each decision node in the quantum decision tree is initialized. A quantum state describes the state of a quantum system, and this initialization process sets the initial conditions for subsequent quantum computations. For example, in quantum computing, the initial state of a qubit might be set to a superposition state—a state in which multiple states exist simultaneously, providing a rich set of possibilities for quantum computation. Through initialization, the quantum system at each decision node enters a state capable of computation and decision-making.

[0136] Step a33: Use quantum bits to encode the input terrain data features and target fusion features to generate the current quantum state.

[0137] Among them, each sub-feature in the target fusion feature corresponds to one or more quantum bits; a quantum bit can represent multiple states at the same time.

[0138] Specifically, quantum bits are used to encode the input terrain data features and target fusion features to generate the current quantum state.

[0139] For example, for each terrain data feature and target fusion feature, according to their mapping relationship, quantum gate operation is used to encode it into the quantum bit. For example, for the elevation feature, the Hadamard gate (H gate) is used to encode it into the superposition state of the quantum bit. When the H gate acts on the |0> state, it is transformed into By adjusting the parameters and action sequence of the H gate, quantum encoding of elevation features is achieved. After encoding a single feature, the interrelationships between features are considered and the qubits of different features are entangled and encoded through quantum gate operations such as the controlled NOT gate (CNOT gate). For example, the qubits of the slope feature and the flow feature are entangled through the CNOT gate, so that the quantum states of the two features are correlated. When the qubit of the slope feature is in the |0> state, the qubit of the flow feature undergoes a corresponding state change based on the action of the CNOT gate, thus achieving quantum encoding of the feature combination.

[0140] Step a34: perform quantum gate operation on the current quantum state to generate node quantum states corresponding to each decision node.

[0141] Specifically, the above step a34 may include the following steps:

[0142] Step a341 , identifying the current quantum state, determining the real-time dynamic features corresponding to the current quantum state, as well as the core features and auxiliary features in the current quantum state.

[0143] Specifically, by measuring quantum states, such as using projection measurements M = {P0, P1} (P0 = |0><0|, P1 = |1><1|), real-time dynamic features are inferred based on the probability distribution of the measurement results. For example, if the probability of the measured |0> state changes significantly over a short period of time, real-time dynamic information such as water flow velocity and precipitation intensity can be determined by combining terrain data with background knowledge of target fusion features.

[0144] Then, we use information-theoretic metrics of quantum states, such as quantum mutual information I(ρ AB )=S(ρ A )+S(ρ B )-S(ρ AB )(ρ A , ρ B (where ρ is the density matrix of subsystems A and B, respectively, and S(ρ) is the von Neumann entropy) is used to evaluate the contribution of different features to the quantum state. For example, in terrain data features and target fusion features, the quantum mutual information between each feature and the overall quantum state is calculated. Features with quantum mutual information greater than a preset threshold are considered core features, as they have a more significant impact on the quantum state. Features with quantum mutual information less than or equal to the preset threshold are considered auxiliary features.

[0145] Step a342: Determine the target quantum gate combination corresponding to the current quantum state based on the real-time dynamic characteristics.

[0146] Specifically, electronic devices can select quantum gates based on real-time dynamic characteristics (such as the rate of change of water flow velocity and the frequency of water level fluctuations) using predefined mapping rules. For example, for characteristics with severe fluctuations, a Hadamard gate (H gate) is selected for state superposition enhancement; for characteristics that require correlation, a controlled NOT gate (CNOT) or Toffoli gate is selected; and for phase-sensitive characteristics, a phase gate (P gate) or T gate is selected.

[0147] Then, the core and auxiliary features of the current quantum state are used to generate a gate sequence. For example, if core features A and B are highly correlated, a CNOT(A,B) operation is generated; if the volatility of feature C needs to be enhanced, an H-CNOT-H sequence is generated.

[0148] Finally, quantum compilation techniques are used to optimize the gate sequence, reduce the number and depth of gates, and generate the target quantum gate combination. For example, adjacent H gates can be canceled, and SWAP gates can be used to adjust the position of qubits to reduce long-distance interactions.

[0149] Step a343: Perform entanglement calculation on the core features based on the controlled NOT gate in the target quantum gate combination to generate high-order entanglement features.

[0150] Specifically, the electronic device selects the quantum bits corresponding to the core features (such as flow rate, slope, etc.) from the real-time dynamic features.

[0151] Then, the controlled NOT gates (CNOT gates) in the target quantum gate combination are used to construct entangled chains or clusters. For example: CNOT(qubit1, qubit2); CNOT(qubit1, qubit3); H(qubit1) to construct a GHZ state; and W state through sequential CNOT and single-bit rotation operations.

[0152] Finally, through entanglement operations, higher-order correlations between core features are encoded into the quantum state, generating higher-order entangled features. For example, the second-order correlation between the flow rate change rate and slope can be extracted by measuring the entanglement degree after CNOT (flow, slope).

[0153] In step a344, the auxiliary features are entangled with the core features based on the controlled NOT gate in the target quantum gate combination to form a star topology structure.

[0154] Specifically, with each core feature as the center, all related auxiliary features are connected through a controlled NOT gate (CNOT gate). For example, core feature A (flow rate) is used as the center, and auxiliary features B (temperature) and C (wind direction) are connected: CNOT(A,B); CNOT(A,C).

[0155] Then, by adjusting the entanglement strength (e.g., using some CNOT gates), the information transfer efficiency is optimized to form a star topology. For example, using a square root CNOT gate Achieve partial entanglement.

[0156] Step a345: Perform a nonlinear transformation on the current quantum state to generate a nonlinear feature corresponding to the current quantum state.

[0157] Specifically, electronic devices can use quantum phase estimation (QPE) or quantum adiabatic evolution to implement nonlinear transformations. For example, phase nonlinearization can be achieved by applying a phase gate (P(θ)) multiple times, or a controlled phase gate (CU) can be used to construct a nonlinear transformation circuit.

[0158] Then, the linear feature is mapped to the nonlinear space. For example, the linear feature x is mapped to sin(x) or x through the quantum circuit. 2 .

[0159] Finally, nonlinear features are extracted through quantum measurements. For example, the projection of the quantum state in different bases is measured to extract nonlinear related information.

[0160] Step a346, based on the high-order entanglement characteristics, star topology structure and nonlinear characteristics, generates the node quantum state corresponding to each decision node.

[0161] Specifically, the electronic device integrates the high-order entanglement features, star topology, and nonlinear features obtained through the above processing through quantum gate operations. For example, multiple control gates (such as MCX gates) are used to combine different features, or quantum Fourier transforms (QFTs) are applied to transform the feature space.

[0162] Then, the integrated features are mapped to the node quantum state |ψ_node> through the unitary transformation U: |ψ_node>=U|ψ_initial>.

[0163] Finally, the generated node quantum state is verified using Quantum State Tomography or fidelity measurement. For example, the fidelity between the generated state and the target state is calculated as F = |<ψ_target|ψ_node>| 2 .

[0164] Step a35, calculating the quantum probability amplitude of the node quantum state corresponding to each decision node.

[0165] Specifically, for the node quantum state |ψ〉=∑_iα_i|i>, α_i is the quantum probability amplitude.

[0166] For small systems containing qubits whose number is less than the threshold, the electronic device directly calculates the probability amplitude through matrix operations. For example: if |ψ>=U|0>, then α_i=<i|U|0> .

[0167] For large systems with qubits greater than or equal to a threshold, the probability amplitude can be estimated by multiple measurements. For example, N measurements of |ψ> are performed and the frequency of state |i> is statistically obtained as f_i≈|α_i|. 2 .

[0168] Among them, the quantity threshold can be 18, 15, or other values. This embodiment does not specifically limit the quantity threshold.

[0169] Step a36, determining a classification path based on each quantum probability amplitude.

[0170] Specifically, for each possible decision path, its cumulative probability is calculated. For example, the probability of path P = {node 1 → node 2 → node 3} is P = |α_1|2·|α_2|2·|α_3|2.

[0171] Then, the cumulative probabilities of each possible decision path are compared, and the decision path with the largest cumulative probability is selected as the classification path. For example: P_max = max(P_1, P_2, ..., P_n).

[0172] Optionally, the maximum cumulative probability is compared with a probability threshold. If the maximum cumulative probability is lower than the probability threshold, recalculation or fusion of multiple paths is triggered. For example, if P_max < 0.6, weighted fusion of the first k paths is considered.

[0173] Step a37: Enter the next level of decision nodes according to each classification path until reaching the leaf node.

[0174] Specifically, according to the selected classification path, the quantum operations of each node are executed in sequence. After passing a node, the quantum state is updated to the output state of the node, and the calculation stops when it reaches a leaf node (no child node).

[0175] In step a38, the leaf node determines the sediment evolution trend category corresponding to the target gate based on the previously calculated quantum probability amplitude and path information.

[0176] Specifically, the leaf node determines the sediment evolution trend category corresponding to the target gate based on the previously calculated quantum probability amplitude and path information.

[0177] For example, a mapping of leaf node states to sediment evolution trend categories is predefined, for example: |00> → sedimentation trend; |01> → scour trend; |10> → equilibrium state; |11> → complex evolution.

[0178] Based on the probability amplitude of the leaf node, calculate the probability of each category. For example: P(deposition) = |α_00|2; P(scour) = |α_01|2. Select the category with the highest probability as the final result. For example: If P(deposition) > P(scour), then the trend is determined to be deposition.

[0179] In step S2032, the sediment evolution trend category, the current scouring and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data are input into a preset sediment scouring change prediction model, and the sediment scouring change of the target gate within the future preset time period is output.

[0180] Specifically, the above step S2032 may include the following steps:

[0181] Step b1: Input sediment evolution trend category, current scouring and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data into the time-dependent network of the preset sediment scouring change prediction model to output the time-dependent features.

[0182] Specifically, the above step b1 may include the following steps:

[0183] In step b11, the sediment evolution trend category, current scouring and deposition terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data are input into the time-dependent network of the preset sediment scouring change prediction model.

[0184] Specifically, the electronic device normalizes heterogeneous data from multiple sources, including sediment evolution trends, current erosion and deposition topography sequences, and future precipitation-runoff data, eliminating dimensional differences. Sediment evolution trends are represented using one-hot encoding or embedding vectors; current erosion and deposition topography sequences are converted into feature vectors using point cloud data processing techniques; and time series data, such as future precipitation-runoff, are normalized. Subsequently, spatiotemporal alignment techniques are used to precisely match these different data across time dimensions, generating structured multimodal input data that is fed into a pre-defined time-dependent network.

[0185] Step b12, the weight determination network in the time-dependent network evaluates the input data at each time step to determine the importance of the input data at each time step to the current erosion and deposition topography sequence and the changes in sediment cross-section erosion within a preset time period in the future.

[0186] Specifically, the structured multimodal input data is combined into an input vector for each time step. Assuming there are n data types, each data type extracts m ifeatures (i=1,2,…,n), then the input vector xt at the tth time step is: xt=[f1,t,f2,t,…,fn,t]; where fi,t represents the feature vector extracted from the i-th data type at the tth time step.

[0187] To make the weight-determining network aware of time information, a time step identifier is introduced. Positional encoding can be used to encode the time step t into a vector pt, which is then concatenated with the input vector xt to obtain the final input representation Xt = [xt; pt]. This allows the network to distinguish data at different time steps and capture the temporal dependencies of the data.

[0188] Then, the input representation Xt is input into the multi-head attention mechanism. The multi-head attention mechanism captures the relationship between data from different perspectives through multiple different "heads". First, Xt is projected into the query, key and value spaces to obtain Q t , K t and V t . Then, for each head h, the attention score is calculated:

[0189]

[0190] Among them, d k is the dimension of the key vector, Q t h , K t h and V t h are the query, key, and value vectors corresponding to the h-th head, respectively. Finally, the results of all heads are concatenated and projected to obtain the attention output At.

[0191] Finally, the attention output At is processed to calculate the preliminary importance score of the input data at each time step. At can be mapped to a scalar value through a linear layer, and then normalized to the interval [0, 1] through an activation function (such as the Sigmoid function) to obtain a preliminary importance score αt, which reflects the relative importance of each time step data in the overall data.

[0192] Step b13: Determine the weight information corresponding to the input data at each time step according to the importance corresponding to the input data at each time step.

[0193] The electronic device determines the weight information corresponding to the input data at each time step according to the relationship between the calculated importance and the weight information.

[0194] Step b14: dynamically adjust the update gate and forget gate in the time-dependent network according to the weight information corresponding to the input data at each time step.

[0195] Specifically, according to the weight information corresponding to the input data at each time step, the parameters of the update gate and the forget gate in the time-dependent network are dynamically adjusted.

[0196] Specifically, if the weight information is greater than the weight threshold, the update gate is opened wider, allowing it to participate more in the state update of the time-dependent network. If the weight information is less than or equal to the weight threshold, the forget gate is closed wider, reducing its impact on network memory. The Q-learning algorithm from reinforcement learning is introduced to continuously optimize the gating strategy and find the gating parameter combination that minimizes the prediction error.

[0197] In step b15, time series features are extracted from the input data according to the update gate and forget gate adjusted at each time step, and time-dependent features are output.

[0198] Specifically, time series features are extracted from the input data using adjusted update and forget gates. Alternatively, a convolutional neural network (CNN) and a recurrent neural network (RNN) can be combined: the CNN extracts local features of the data, while the RNN captures long-term dependencies in the time series. A self-attention mechanism is then used to weight the extracted features, highlighting key features and outputting time-dependent features.

[0199] In step b2, the sediment evolution trend category, the current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate operation data are input into the structural feature extraction network in the preset sediment scouring change prediction model to output the topological structure features.

[0200] Specifically, the above step b2 may include the following steps:

[0201] Step b21: construct a target graph structure based on the current scouring and silting terrain sequence, future precipitation-runoff data, future upstream and downstream water level difference, future tidal level change data, and future gate scheduling data.

[0202] Specifically, electronic devices can convert various types of data into an initial graph structure G = (V, E), where: Node V: contains hydrological monitoring points (such as upstream and downstream water level stations, gate locations), terrain feature points (key points of riverbed sections), meteorological monitoring stations, etc. Each node contains a multidimensional feature vector (such as water level, flow, terrain elevation). Edge E: represents the physical relationship between nodes and is constructed in three ways, namely: Spatial association edge: constructed based on geographical location distance, such as the weight of the connection edge between adjacent water level stations is inversely proportional to the distance. Time-dependent edge: Directed edges are constructed between nodes at different time steps of the same monitoring point to form a time series chain. Physical causal edge: constructed based on the principles of hydrodynamics, such as the downstream water level is affected by the upstream flow and gate scheduling, and the direction and weight of the edge are determined by a causal reasoning algorithm (such as the PC algorithm).

[0203] Then, based on real-time hydrological data, the attention mechanism adjusts edge weights to generate the target graph structure. For example, during flood season, edge weights between upstream and downstream water level stations are automatically enhanced. Gated recurrent units (GRUs) are used to aggregate historical information, so that node features contain temporal evolution information:

[0204]

[0205] Where ht is the hidden state of node u at time t, and N(u) is the set of neighbor nodes.

[0206] Step b22: extract features from the graph structure based on the structural feature extraction network and output topological structure features.

[0207] Specifically, the network is extracted based on structural features to capture the spatial dependencies between nodes:

[0208] in Add self-loops to the adjacency matrix, is the degree matrix, W (l) is a learnable weight. Then, one-dimensional convolution and attention mechanism are combined to capture temporal features: t =Attention(H t , H t-1 ,...,H t-k ), where k is the time window size, and the attention mechanism automatically assigns weights to different time steps.

[0209] The spatial dependency and temporal features are integrated to output topological structure features.

[0210] In step b3, the time-dependent features and topological structure features are fused and the fused prediction features are output.

[0211] Specifically, the time-dependent features and topological structure features are mapped to the same dimensional space through linear transformation: FT′=WTFT, FG′=WGFG, where FT is the time-dependent feature and FG is the topological structure feature.

[0212] Then, the attention score of the time-dependent features to the topological structure features is calculated: Where [;] represents the splicing operation, W a are learnable weights.

[0213] Then, based on the attention score of the time-dependent feature on the topological structure feature, the transformed time-dependent feature and the topological structure feature are fused to output the fused prediction feature. The specific formula is as follows:

[0214]

[0215] in Represents element-wise multiplication.

[0216] Step b4: Based on the integrated prediction features, output the sediment cross-section changes corresponding to the target gate within a preset time period in the future.

[0217] Specifically, the preset sediment impact change prediction model can take into account the target physical constraints and output the sediment cross-section impact change within a preset time period in the future corresponding to the target gate based on the integrated prediction characteristics.

[0218] The target physical constraints include minimizing the mean square error between the predicted value and the observed value: Among them, y pred is the predicted value, y obrs is the observed value in the current erosion and deposition terrain series; ensure that the predicted changes of adjacent time steps are smooth: Design constraints based on the sediment transport equation: Where S is the sediment concentration, q is the water velocity, D and E are the sedimentation and scour rates respectively.

[0219] The embodiment of the present application provides a method for predicting changes in sediment cross-section scouring under the influence of a sluice gate, which obtains historical precipitation-runoff data, historical upstream and downstream water level differences, historical tidal level change data, and historical gate scheduling data corresponding to historical precipitation-runoff data, historical upstream and downstream water level differences, and historical tidal level change data, comprehensively covering the key factors affecting sediment scouring and deposition. It provides a rich and accurate data basis for subsequent analysis and avoids analytical deviations caused by missing data. Based on historical precipitation-runoff data, historical upstream and downstream water level differences, historical tidal level change data, and historical gate scheduling data, a causal graph structure is constructed, which can clearly and intuitively present the causal relationship between various factors, help understand the inherent mechanism in the sediment scouring and deposition process, make the model more interpretable, and facilitate water conservancy experts to verify and optimize the model logic. The causal graph is expanded into a dynamic structural equation model to capture the time-lag causal relationship between variables, taking into account the time delay effect in the hydrological process, and improving the accuracy and reliability of model predictions. The future precipitation-runoff data, the future upstream and downstream water level difference and the future tidal level change data are substituted into the time-lag causal relationship to determine the future gate scheduling data corresponding to the target gate, thereby ensuring the accuracy of the determined future gate scheduling data.

[0220] Then, the current scouring and silting terrain sequence, future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate scheduling data are input into the scouring and silting trend classification model. Feature extraction is performed on the current scouring and silting terrain sequence to obtain the terrain data features corresponding to the current scouring and silting terrain sequence, ensuring the accuracy of the terrain data features corresponding to the current scouring and silting terrain sequence. The chaotic characteristics of future data are analyzed using a preset adaptive phase space reconstruction algorithm, which can explore the complex nonlinear change laws in the hydrological system, capture subtle change trends that are difficult to detect with traditional methods, and provide more in-depth information for prediction. The chaotic features are input into the reinforcement learning model to determine the target fusion strategy, avoiding the subjectivity and limitations of artificially set fusion rules, making the fused features more representative and effective. The target fusion features are generated based on the target fusion strategy, integrating various aspects of chaotic information, enhancing the feature's ability to express sediment evolution trends, and providing better input for subsequent classification and prediction.

[0221] Next, the terrain data features and target fusion features are input into the quantum decision tree in the scouring and silting trend classification model. The introduction of the quantum decision tree utilizes the superposition and entanglement characteristics of the quantum system. Compared with the traditional decision tree, it can process multiple possibilities at the same time, greatly improving the computing efficiency and the ability to process complex data. It is especially suitable for highly complex scenarios such as hydrological systems. From quantum state initialization, feature encoding to quantum gate operations, the input features are deeply processed. By generating high-order entangled features, constructing star topology structures and nonlinear features, the complex relationship between features is fully explored, and the accuracy of classification and sensitivity to subtle changes are further improved. The quantum probability amplitude is calculated and the classification path is determined. The sediment evolution trend category is determined in a probabilistic form. Not only the classification results are given, but also the credibility information of the results is provided, providing a more comprehensive reference for decision-making.

[0222] After determining the sediment evolution trend category corresponding to the target gate, the sediment evolution trend category, the current scouring and deposition topographic sequence, future precipitation-runoff data, future upstream and downstream water level differences, future tidal level change data, and future gate scheduling data are input into the time-dependent network of the preset sediment scouring and deposition change prediction model. This comprehensively considers the impact of multiple factors on sediment scouring and deposition, avoids the one-sidedness of single-factor analysis, and makes the prediction more consistent with actual conditions.

[0223] The weight determination network assigns weights based on data importance, adaptively highlighting the role of key factors at different time steps. For example, during flood season, the weight of precipitation-runoff data is increased, improving the model's adaptability to diverse hydrological conditions and forecasting accuracy. Dynamically adjusting the update and forget gates effectively controls the inflow and outflow of information, extracting more valuable time series features, enhancing the model's ability to capture temporal dependencies in the data, and improving forecasting accuracy and stability. Using the adjusted update and forget gates at each time step, time series features are extracted from the input data and output as time-dependent features, ensuring the accuracy of the output time-dependent features.

[0224] Next, a target graph structure is constructed based on the current scouring and deposition topography sequence, future precipitation-runoff data, future upstream and downstream water level differences, future tidal changes, and future gate operation data. This better reflects the topological structure and flow propagation patterns of the river system, providing richer structural information for prediction. The network outputs topological structure features using structural features, further exploring the hidden information within the graph structure and complementing the time-dependent features. Finally, the time-dependent and topological structure features are fused to produce a fused prediction feature. Based on this fused prediction feature, the sediment cross-section scouring and deposition changes corresponding to the target gate within a preset time period are output. This organically integrates spatiotemporal information, enabling the model to comprehensively consider the impact of temporal evolution and spatial structure on sediment scouring and deposition. Ultimately, it produces more accurate predictions of future sediment cross-section scouring and deposition changes, providing a scientific and reliable basis for water conservancy project decision-making, flood control scheduling, and other aspects.

[0225] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for predicting changes in sediment cross-section under the influence of a sluice gate, characterized in that: The method comprises: Obtain the current scouring and silting terrain sequence corresponding to the target gate at the current moment, and obtain future precipitation-runoff data, future upstream and downstream water level difference, and future tidal level change data corresponding to the target gate within a preset future time period; Determining future gate scheduling data corresponding to the target gate based on the future precipitation-runoff data, the future upstream and downstream water level difference, and the future tidal level change data; The current scouring and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data and the future gate scheduling data are input into the preset scouring and deposition change prediction model, and the scouring and deposition change of the sediment section corresponding to the target gate within the future preset time period is output.

2. The method according to claim 1, characterized in that The preset scour change prediction model includes a scour and deposition trend classification model and a sediment scour change prediction model; the current scour and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data are input into the preset scour change prediction model, and the output of the sediment cross-section scour change corresponding to the target gate within the future preset time period includes: Inputting the current scouring and silting topography sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data into the scouring and silting trend classification model to determine the sediment evolution trend category corresponding to the target gate; The sediment evolution trend category, the current scouring and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data and the future gate scheduling data are input into the preset sediment scouring change prediction model, and the sediment cross-section scouring change within the future preset time period corresponding to the target gate is output.

3. The method according to claim 2, characterized in that The current scouring and deposition topography sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data are input into the scouring and deposition trend classification model to determine the sediment evolution trend category corresponding to the target gate, including: Performing feature extraction on the current scouring and silting terrain sequence to obtain terrain data features corresponding to the current scouring and silting terrain sequence; fusing the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data to generate target fusion features; The terrain data features and the target fusion features are input into the scouring and silting trend classification model to determine the sediment evolution trend category corresponding to the target gate.

4. The method according to claim 3, characterized in that The fusing of the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data to generate target fusion features includes: Based on a preset adaptive phase space reconstruction algorithm, analyzing the chaotic characteristics corresponding to the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data; the chaotic characteristics include the maximum Lyapunov exponent, fractal dimension, and chaotic frequency spectrum; Inputting the chaotic features corresponding to the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data into the reinforcement learning model in the scouring and deposition trend classification model, and determining the target fusion strategy corresponding to each chaotic feature; The chaotic features are fused based on the target fusion strategy to generate a target fusion feature.

5. The method according to claim 3, characterized in that Inputting the terrain data features and the target fusion features into the scouring and silting trend classification model to determine the sediment evolution trend category corresponding to the target gate includes: Inputting the terrain data features and the target fusion features into the quantum decision tree in the scouring and silting trend classification model; each decision node in the quantum decision tree is an independent quantum system; Initializing the quantum state of each decision node in the quantum decision tree; The input terrain data features and the target fusion features are encoded using quantum bits to generate a current quantum state; each sub-feature in the target fusion feature corresponds to one or more quantum bits; and one quantum bit can represent multiple states simultaneously; Performing quantum gate operations on the current quantum state to generate node quantum states corresponding to the decision nodes; Calculating the quantum probability amplitude of the node quantum state corresponding to each of the decision nodes; determining a classification path based on each of the quantum probability amplitudes; According to each of the classification paths, enter the next layer of decision nodes until reaching the leaf node; The leaf node determines the sediment evolution trend category corresponding to the target gate based on the previously calculated quantum probability amplitude and path information.

6. The method according to claim 5, characterized in that The performing a quantum gate operation on the current quantum state to generate a node quantum state corresponding to each decision node includes: Identifying the current quantum state, determining real-time dynamic features corresponding to the current quantum state, and core features and auxiliary features in the current quantum state; Determining a target quantum gate combination corresponding to the current quantum state based on the real-time dynamic characteristics; performing entanglement calculation on the core feature based on the controlled NOT gate in the target quantum gate combination to generate a high-order entanglement feature; Entangling the auxiliary feature with the core feature based on a controlled NOT gate in the target quantum gate combination to form a star topology structure; Performing a nonlinear transformation on the current quantum state to generate a nonlinear feature corresponding to the current quantum state; Based on the high-order entanglement characteristics, the star topology and the nonlinear characteristics, the node quantum state corresponding to each decision node is generated.

7. The method according to claim 2, characterized in that The step of inputting the sediment evolution trend category, the current scouring and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data into a preset sediment scouring change prediction model, and outputting the sediment cross-sectional scouring change corresponding to the target gate within the future preset time period, includes: Inputting the sediment evolution trend category, the current scouring and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate operation data into the time-dependent network of the preset sediment scouring and deposition change prediction model, and outputting the time-dependent features; Inputting the sediment evolution trend category, the current scouring and deposition topography sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate operation data into the structural feature extraction network in the preset sediment scouring and deposition change prediction model, and outputting the topological structure features; Fusing the time-dependent features and the topological structure features to output a fused prediction feature; Based on the integrated prediction features, the sediment cross-section changes within the future preset time period corresponding to the target gate are output.

8. The method according to claim 7, characterized in that The sediment evolution trend category, the current scouring and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data are input into the time-dependent network of the preset sediment scouring and deposition change prediction model, and the time-dependent features are output, including: Inputting the sediment evolution trend category, the current scouring and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate operation data into the time-dependent network of the preset sediment scouring and deposition change prediction model; The weight determination network in the time-dependent network evaluates the input data at each time step to determine the importance of the input data at each time step to the current erosion and deposition topography sequence and the erosion and deposition changes of the sediment section within the future preset time period; Determining weight information corresponding to the input data at each time step according to the importance corresponding to the input data at each time step; Dynamically adjust the update gate and the forget gate in the time-dependent network according to the weight information corresponding to the input data at each time step; According to the update gate and forget gate adjusted at each time step, time series feature extraction is performed on the input data, and the time-dependent feature is output.

9. The method according to claim 7, characterized in that The structural feature extraction network of the preset sediment erosion change prediction model is input with the sediment evolution trend category, the current scouring and deposition topography sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data, and outputs topological structural features, including: Constructing a target graph structure based on the current scouring and deposition terrain sequence, the future precipitation-runoff data, the future upstream and downstream water level difference, the future tidal level change data, and the future gate scheduling data; The graph structure is subjected to feature extraction based on the structural feature extraction network, and the topological structure features are output.

10. The method according to claim 1, characterized in that The determining of future gate dispatching data corresponding to the target gate based on the future precipitation-runoff data, the future upstream and downstream water level difference, and the future tidal level change data includes: Obtaining historical precipitation-runoff data, historical upstream and downstream water level differences, historical tidal level change data, and historical gate dispatching data corresponding to the historical precipitation-runoff data, the historical upstream and downstream water level differences, and the historical tidal level change data; Constructing a causal graph structure based on the historical precipitation-runoff data, the historical upstream and downstream water level difference, the historical tide level change data, and the historical gate scheduling data; The causal graph structure is extended to a dynamic structural equation model to capture the time-lagged causal relationship between variables; Substitute the future precipitation-runoff data, the future upstream and downstream water level difference and the future tide level change data into the time-lag causal relationship to determine the future gate scheduling data corresponding to the target gate.

Images