Reservoir group flood control reserved storage capacity real-time optimization distribution method considering random incoming water

Abstract

The invention discloses a real-time optimal distribution method for flood control reserved storage capacity of a reservoir group considering random incoming water. The method comprises the following steps: generating a reservoir optimal scheduling first sample set oriented to random simulated flood based on a reservoir flood control rule optimal scheduling model; defining and extracting flood control dispatching benefit curve clusters around the first dispatching sample set, and correcting the flood control dispatching benefit curve clusters to obtain a second dispatching sample set oriented to random simulation flood; identifying a key flood characteristic factor set, a downstream peak clipping amount, an upstream reservoir input capacity and other characteristics in the second scheduling sample set, fitting a mapping relation among the sample characteristics by adopting a machine learning method to obtain a flood control benefit function, and taking the flood control benefit function as a constraint condition of a real-time optimization distribution model; and meanwhile, constructing a reservoir group reserved flood control capacity real-time optimal distribution model by taking the maximum residual flood control capacity of the reservoir group as a target, designing a model traversal solving strategy, obtaining a reservoir capacity optimal distribution scheme of the reservoir group according to different peak clipping flow / water level targets, and calculating reliability verification evaluation of the scheme.

Description

Technical Field

[0001] This application relates to the field of reservoir flood control technology, and in particular to a real-time optimization allocation method for the flood control reserved capacity of a reservoir group that takes into account random water inflow. Background Technology

[0002] During floods, upstream reservoirs can effectively reduce excess floodwater at downstream flood control points, lower water levels in key river sections, and mitigate flood losses in downstream flood-prone areas through various measures such as pre-release of water to create storage capacity and flood peak reduction. With the construction and operation of large, controlling upstream reservoirs, the flood control capacity of the basin has been greatly enhanced. However, the flood control pattern in different regions has also changed significantly. Different upstream reservoirs have varying flood control capabilities, resulting in different flood compensation capacities for downstream control points. Furthermore, the flood control benefits of different upstream reservoirs vary depending on the composition of the incoming water from different regions.

[0003] Therefore, how to analyze the flood control scheduling benefits of reservoir groups under changing scenarios and provide flood control storage capacity allocation schemes for different water inflow scenarios, so as to further support the optimization and formulation of flood control scheduling rules and real-time scheduling schemes for reservoir groups, has become a cutting-edge research topic in the field of scheduling. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides a real-time optimization allocation method for flood control reserve capacity of reservoir groups that considers random inflows. This method can solve the problem that the rationality of the flood control capacity optimization allocation scheme is low due to the single design flood process, and the calculation results are only used to guide the reservoir planning and design stage and are difficult to use in real-time flood control scheduling.

[0005] Firstly, this application provides a real-time optimized allocation method for the flood control reserve capacity of a reservoir group considering random inflows, which includes: Based on the pre-set reservoir flood control rule optimization scheduling model, the first scheduling sample set of reservoir optimization scheduling for random simulated floods is generated. Based on the first scheduling sample set, a cluster of flood control scheduling benefit curves is defined and extracted, and the cluster of flood control scheduling benefit curves is modified to obtain a second scheduling sample set for random simulated floods; The sample features in the second scheduling sample set are identified, and the mapping relationship between the sample features is fitted by machine learning method to obtain the flood control benefit function of the reservoir for the downstream flood control target under the scenario of random water inflow. The sample features include the key flood feature factor set, the downstream peak reduction amount and the upstream reservoir input capacity. Using the flood control benefit function as a constraint condition for the real-time optimization allocation model, a real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the goal of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed to obtain the reservoir group's capacity optimization allocation scheme based on different peak flow or water level targets, and the reliability of the calculation scheme is verified and evaluated.

[0006] Secondly, this application provides a real-time optimized allocation device for the flood control reserve capacity of a reservoir group that takes into account random inflows, comprising: The first generation unit is used to generate a first scheduling sample set for reservoir optimization scheduling oriented to random simulated floods, based on a pre-set reservoir flood control rule optimization scheduling model. The correction unit is used to define and extract a cluster of flood control scheduling benefit curves around the first scheduling sample set, and to correct the cluster of flood control scheduling benefit curves to obtain a second scheduling sample set for random simulated floods. The fitting unit is used to identify sample features in the second scheduling sample set and to fit the mapping relationship between each sample feature using machine learning methods to obtain the flood control benefit function of the reservoir for the downstream flood control target under the scenario of random water inflow. The sample features include the key flood feature factor set, downstream peak reduction amount and upstream reservoir input capacity. The second generation unit is used to use the flood control benefit function as a constraint condition for the real-time optimization allocation model. With the goal of maximizing the remaining flood control capacity of the reservoir group, a real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed. The model traversal solution strategy is designed to obtain the reservoir group's capacity optimization allocation scheme according to different peak flow or water level targets, and the reliability of the calculation scheme is verified and evaluated.

[0007] Thirdly, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the real-time optimized allocation method for flood control reserved capacity of a reservoir group considering random water inflows as provided in the first aspect above.

[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to execute the real-time optimized allocation method for flood control reserved capacity of a reservoir group considering random water inflows provided in the first aspect.

[0009] Fifthly, embodiments of this application also provide a computer program product, including a computer program or instructions, which are executed by a processor to provide a real-time optimized allocation method for flood control reserved capacity of a reservoir group considering random water inflows, as provided in the first aspect.

[0010] The real-time optimization allocation method for flood control reserved capacity of reservoir groups considering random inflows provided in this application embodiment generates a first sample set of reservoir optimization scheduling for random simulated floods based on a pre-set reservoir flood control rule optimization scheduling model. Around the first scheduling sample set, a cluster of extracted flood control scheduling benefit curves is defined and modified to obtain a second scheduling sample set for random simulated floods. Key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input in the second scheduling sample set are identified. Machine learning methods are used to fit the mapping relationship between each sample feature to obtain the flood control benefit function of the reservoir for downstream flood control objectives under the scenario of random inflows. The objective flood control benefit function is used as a constraint condition for the real-time optimization allocation model. A real-time optimization allocation model for reserved flood control capacity of the reservoir group is constructed with the objective of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed, and then the reservoir group capacity optimization allocation scheme is obtained according to different peak reduction flow / water level objectives. The reliability of the calculation scheme is verified and evaluated. The proposed method solves the problem of low rationality of flood control capacity optimization allocation schemes caused by a single design flood process. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating the real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random water inflow, provided in an embodiment of this application. Figure 2 This is a schematic diagram of the scheduling sample correction strategy in the second scheduling sample set provided in the embodiments of this application; Figure 3 A schematic block diagram of a real-time optimized allocation device for flood control reserve capacity of a reservoir group that takes into account random water inflow, provided in an embodiment of this application. Figure 4 A schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0015] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0016] It should also be further understood that the term “and / or” as used in this application specification and the appended claims can be understood as any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0017] Furthermore, in this application, unless otherwise explicitly specified or limited in the embodiments, the terms "installation," "connection," "joining," and "fixing" appearing in the embodiments should be interpreted broadly. For example, a connection can be a fixed connection, a detachable connection, or an integral part; it can also be a mechanical connection, an electrical connection, etc. Of course, it can also be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication between two components, or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific implementation.

[0018] In related technologies, studies on the analysis of flood control scheduling benefits and reservoir capacity allocation for reservoir groups mostly use historical measured sequences or typical overall design floods as model inputs for analysis. However, the length of measured data sequences is often short, the characteristic values ​​of each flood in the measured years are difficult to reach the specified frequency, and the number of typical design flood years is even smaller, the composition of flood regions is relatively homogeneous, and the data production process is often based on basic assumptions such as floods in different zones having the same frequency and flood volumes having the same frequency in different time periods, resulting in an unclear concept of flood frequency, which in turn affects the reliability of the scheduling benefit analysis and reservoir capacity allocation results.

[0019] To this end, this application provides a real-time optimization allocation method for the flood control reserved capacity of a reservoir group that considers random water inflow. This method is used to analyze the flood control scheduling effect of the reservoir group under random water inflow scenarios and quickly provide a flood control capacity allocation scheme for the reservoir group. Specifically, based on a pre-defined reservoir flood control rule optimization scheduling model, a first sample set of reservoir optimization scheduling for random simulated floods is generated. Around this first scheduling sample set, a cluster of flood control scheduling benefit curves is defined and modified to obtain a second scheduling sample set for random simulated floods. Key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input are identified within the second scheduling sample set. Machine learning methods are used to fit the mapping relationships between these sample features, yielding a flood control benefit function for the reservoir's downstream flood control objective under random inflow scenarios. This objective flood control benefit function is used as a constraint on a real-time optimization allocation model. A real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the objective of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed, and then, based on different peak reduction flow / water level objectives, the reservoir group's capacity optimization allocation scheme is obtained. The reliability of the calculation scheme is verified and evaluated. The proposed method solves the problem of low rationality in flood control capacity optimization allocation schemes caused by a single design flood process.

[0020] It is understood that the real-time optimized allocation method for flood control reserve capacity of reservoir groups considering random water inflows provided in this application embodiment can be applied to terminal devices. This method uses application software installed on the terminal device to identify sudden shifts between flood and drought events. The terminal device can be a desktop computer, laptop computer, tablet computer, or mobile phone, etc.

[0021] It should be noted that the application scenarios described in the following embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0022] The following is a detailed description of the real-time optimization allocation method for flood control reserve capacity of reservoir groups that takes into account random water inflow, provided in this application.

[0023] like Figure 1 As shown, the method includes the following steps S110 to S140.

[0024] S110. Based on the pre-set reservoir flood control rule optimization scheduling model, generate the first scheduling sample set for reservoir optimization scheduling oriented to random simulated floods.

[0025] In this embodiment, the reservoir flood control rule optimization scheduling model can be a pre-constructed model for optimizing reservoir flood control scheduling. This application can input key feature information of randomly simulated floods into the reservoir flood control rule optimization scheduling model and perform optimization scheduling through the model, thereby generating multiple first scheduling samples to form a first scheduling sample set. The first scheduling sample set includes multiple first scheduling samples.

[0026] In some embodiments, before generating the first scheduling sample set of random simulated floods based on a pre-set reservoir flood control rule optimization scheduling model, the method further includes: determining the scheduling rules for upstream reservoir flood control; determining, according to the scheduling rules, the scheduling discrimination index of the upstream reservoir for downstream flood control control points, the corresponding scheduling method of the upstream reservoir, and the classification method of the scheduling discrimination index; the scheduling discrimination index includes at least one of the following: real-time inflow index of the reservoir, forecasted rainfall index, real-time and forecasted water level and flow index of the downstream flood control control point, and current water level of the reservoir; the scheduling method includes equal storage and compensation scheduling; and constructing a reservoir flood control rule optimization scheduling model based on the scheduling discrimination index, classification method, and scheduling method.

[0027] Specifically, in the process of constructing the reservoir flood control rule optimization scheduling model, this application can be based on the reservoir flood control scheduling rules. Specifically, the scheduling rules of the upstream reservoir flood control can be determined in advance. Then, parameterized modeling and optimization solution can be performed based on the reservoir forecast scheduling rules. In the process of parameterized modeling and optimization solution based on the reservoir forecast scheduling rules, the scheduling discrimination index and scheduling mode determination grading method of the upstream reservoir can be determined in advance. After that, the objective constraints are determined and optimization solution is performed.

[0028] The scheduling criteria include at least one of the following: real-time inflow to the reservoir, forecasted rainfall over the corresponding period, real-time and forecasted water level and flow at downstream flood control points, and the current water level of the reservoir. Scheduling methods include equal-capacity impoundment and compensatory scheduling. Specifically, the scheduling criteria for upstream reservoirs generally include real-time inflow to the reservoir, forecasted rainfall over the corresponding period, real-time and forecasted water level and flow at downstream flood control points, and the current water level of the reservoir. When the reservoir has a large flood control capacity and the flood forecast accuracy is high, if there is an important flood control area downstream, the forecasted flow or water level at the downstream flood control points is generally chosen as the key scheduling criteria. Reservoirs generally use equal-capacity impoundment and compensatory scheduling methods to manage downstream flood control points.

[0029] Equal storage capacity retention refers to using a fixed retention capacity to retain downstream flood peaks for different water inflow processes. This scheduling method is generally used when the reservoir is far from the flood control station. In this case, the parameter to be optimized is the size of the fixed retention capacity of the reservoir.

[0030] Compensatory flood control refers to a reservoir's real-time interception and storage of excess floodwater at downstream flood control points based on forecast information. This method is often used when the reservoir has strong flood control capacity and is close to the flood control station. In this case, the parameters to be optimized can be the ratio of excess floodwater at the downstream flood control point to the intercepted flow. .

[0031] In some embodiments, a reservoir flood control rule optimization scheduling model is constructed based on scheduling discrimination indicators, classification methods, and scheduling methods. This includes: classifying the scheduling discrimination indicators, parameterizing the reservoir scheduling methods under different classification discrimination indicators, determining the flood control rule optimization scheduling objective, and using an optimization algorithm to solve for the classification discrimination indicators and scheduling method modeling parameters.

[0032] In this embodiment, the proposed indicators are classified according to scheduling experience. Generally, the number of classifications should not be excessive, and the decision space after classification must form a complete closed loop to encompass all possible decision scenarios. Table 1 divides each indicator into m levels. and These represent the values ​​of the j-th type discrimination index to be optimized at level k in the pre-accumulation and pre-release stages, respectively.

[0033] The classification description of the flood control dispatching rules indicators can be found in Table 1: Table 1. Classification Description of Flood Control Dispatch Rules Indicators

[0034] In some embodiments, the objective function of the reservoir flood control rule optimization scheduling model includes the cumulative remaining flood control capacity of the upstream reservoir; the constraints of the reservoir flood control rule optimization scheduling model include at least one of the following: water level and flow constraints at the downstream flood control control point, water balance constraints at the inflow and outflow of the upstream reservoir, and maximum and minimum constraints on water level and flow.

[0035] In this embodiment, during the determination of the objective constraints, the maximum flood control safety guarantee criterion can be considered, that is, the reservoir uses as little storage capacity as possible while ensuring safe downstream flow to prepare for the next possible flood. Its scheduling objective function is as follows:

[0036] In the formula, It is the total flood control capacity of the reservoir; The cumulative reservoir capacity consumed at time t; T is the length of the scheduling period; To accumulate the remaining flood control capacity of the reservoir.

[0037] The constraints during the scheduling process are analyzed based on specific circumstances, generally including water level and flow constraints at downstream flood control points, inflow and outflow balance constraints at upstream reservoirs, and maximum and minimum constraints on water level and flow. The water level and flow constraints at downstream flood control points are shown below:

[0038] In the formula, Let t be the outflow from the reservoir during time period t, calculated as the flow rate to the downstream flood control point j via the river channel. The inflow between the upstream reservoir group and flood control station i during time period t. This is the highest water level at the control station after reservoir scheduling; , The safe flow rate and safe water level for flood control station j.

[0039] During the model solving process, the initial range of each parameter to be optimized can be determined, and each index can be optimized according to different water inflow processes. If the number of indexes is small, grid search can be used; if the number of indexes is too large, heuristic intelligent algorithms such as particle swarm optimization or ant colony optimization can be used for solving the problem.

[0040] In some embodiments, a first scheduling sample set for random simulated floods is generated based on a pre-set reservoir flood control rule optimization scheduling model. This includes: using a pre-set multi-station flood simulation model to randomly simulate the flood process to obtain key feature information of the simulated flood; the key feature information includes the upstream reservoir flood peak, the downstream flood control point flood peak, and the flood peak lag time; based on the key feature information, the reservoir flood control rule optimization scheduling model is used to optimize the scheduling of the upstream reservoir to obtain the upstream reservoir's input capacity value, the downstream flood control point's peak reduction flow, and the maximum water level reduction value; and combining the key feature information, the upstream reservoir's input capacity value, the downstream flood control point's peak reduction flow, and the maximum water level reduction value to obtain the first scheduling sample set.

[0041] In this embodiment, the reservoir flood control rule optimization scheduling model can perform stochastic simulation of flood processes. It can refer to the multi-station flood process stochastic simulation method considering historical extreme flood characteristics in CN113946940A to perform stochastic simulation of flood processes, thereby obtaining a multi-station simulated flood sequence considering historical flood characteristics. Then, feature extraction is performed to obtain key feature information of the simulated flood, such as peak flow at different stations, maximum flood volume over 7 days, 15 days, and 30 days, and peak delay. Based on this key feature information, the reservoir flood control rule optimization scheduling model is used to optimize the scheduling of the upstream reservoir, obtaining the upstream reservoir's input capacity, the peak reduction flow at the downstream flood control point, and the maximum water level reduction. The key feature information, the upstream reservoir's input capacity, and the downstream flood control point's peak reduction flow and maximum water level reduction are combined to obtain the first scheduling sample set. The key feature information includes the upstream reservoir peak flow, the downstream flood control point peak flow, and the peak delay.

[0042] Specifically, this application, based on key feature information, uses a reservoir flood control rule optimization scheduling model to optimize the scheduling of upstream reservoirs, obtaining the upstream reservoir's capacity input, the peak reduction flow at downstream flood control points, and the reduction in maximum water level. Random inflow events can be input into the reservoir flood control rule optimization scheduling model. For each inflow event, the safe water level or safe flow at the downstream flood control point is discretely changed, and optimization is performed to obtain the water level and flow process after scheduling of the reservoir and downstream flood control points. Furthermore, the capacity input of the upstream reservoir and the peak reduction flow and reduction in maximum water level at the downstream control points are statistically analyzed for that inflow event. The first scheduling sample set M is as follows:

[0043] S120. Based on the first scheduling sample set, define and extract the flood control scheduling benefit curve cluster, and modify the flood control scheduling benefit curve cluster to obtain the second scheduling sample set for random simulated floods.

[0044] Specifically, this application can extract and correct a cluster of flood control scheduling benefit curves based on the first scheduling sample set to avoid unreasonable results from subsequent multi-input single-output fitting using deep learning methods. Specifically, this application can define the reservoir flood control benefit curve as the relationship between the upstream reservoir's input capacity and the maximum peak-shaving flow or water level at the existing downstream control station, as shown in the attached figure. Figure 2 The first scheduling sample set can be viewed as a combination of different flood characteristic values ​​and their corresponding flood control benefit functions. In the process of using this dataset, it can be regarded as finding the reservoir capacity under the target peak reduction condition by screening flood characteristics and then based on the relationship of flood control benefit functions. Therefore, if a deep learning method is used to fit the sample set with multiple inputs and single outputs, it is necessary to specify the upper and lower limits of the reservoir capacity to be put into each reservoir to avoid unreasonable situations being derived by extrapolating from the curve.

[0045] In some embodiments, the flood control scheduling benefit curve cluster is modified to obtain a second scheduling sample set for random simulated floods, including: formulating a modification strategy for modifying the flood control scheduling benefit curve cluster; and inserting discrete points into the flood control scheduling benefit curve cluster based on the modification strategy to modify the flood control scheduling benefit curve cluster and obtain a second scheduling sample set for random simulated floods.

[0046] In this application, a correction strategy for the flood control scheduling benefit curve cluster can be formulated. Then, based on the correction strategy, discrete points are inserted into the flood control scheduling benefit curve cluster to correct the flood control scheduling benefit curve cluster, thereby obtaining a second scheduling sample set for random simulated floods.

[0047] Specifically, this application can extract the upper and lower limits of reservoir capacity under different flood events, based on the flood control benefit function corresponding to those flood events. Then, discrete point samples are inserted along the coordinate axis for correction, and combined with flood characteristic values ​​to form a new scheduling sample set, i.e., the second scheduling sample set. The specific correction strategy is detailed in the appendix. Figure 2 Specifically, discrete points can be inserted at points A and B to avoid unreasonable extrapolation.

[0048] S130. Identify the sample features in the second scheduling sample set, and use machine learning methods to fit the mapping relationship between each sample to obtain the flood control benefit function of the reservoir for the downstream flood control target under the scenario of random water inflow; the sample features include the key flood feature factor set, the downstream peak reduction amount and the upstream reservoir input capacity.

[0049] In this embodiment, the complex hydraulic response relationship is transformed into a computable mathematical constraint by establishing a flood control benefit function, which can improve the solution efficiency of the optimization model. The multi-reservoir collaborative optimization mechanism overcomes the problem of overall benefit loss caused by local optimization in traditional methods, maximizing the utilization rate of flood control resources of the reservoir group while ensuring downstream flood control safety. The sample features include a set of key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input.

[0050] In some embodiments, identifying sample features in the second scheduling sample set and using machine learning methods to fit the mapping relationships between samples to obtain a flood control benefit function of the reservoir for downstream flood control objectives under a random inflow scenario includes: using correlation coefficient analysis or grey relational analysis to analyze and filter the second scheduling sample set to construct sample set data; and using machine learning methods to fit the mapping relationships between the features of each sample in the sample set data to obtain a flood control benefit function of the reservoir for downstream flood control objectives under a random inflow scenario.

[0051] In this embodiment, during the process of identifying the sample features in the second scheduling sample set, a sample set data covering key flood feature vectors, upstream reservoir capacity, and downstream flood control point peak reduction can be constructed through feature extraction.

[0052] Specifically, this application uses factor analysis methods such as correlation coefficient analysis and grey relational analysis to analyze the first scheduling sample set, identify and extract key flood characteristics that are highly correlated with reservoir capacity input, establish set-pair relationships between key flood characteristics, downstream peak reduction, and upstream reservoir capacity, and construct the sample set data. The sample set data structure is shown in Table 2 below.

[0053] Table 2 Sample Set Data Structure

[0054] In some embodiments, correlation coefficient analysis or grey relational analysis is used to analyze and screen the second scheduling sample set to construct sample set data, including: using correlation coefficient analysis or grey relational analysis to analyze and screen the second scheduling sample set to identify and extract key flood features that are related to the reservoir's capacity input; and establishing set pair relationships between key flood features, downstream peak reduction, and upstream reservoir capacity to construct sample set data.

[0055] In this embodiment, correlation coefficient analysis can be understood as quantifying the linear or monotonic correlation between the upstream reservoir's capacity input and the downstream flood control point's peak flow and water level reduction in the scheduling sample by calculating the Pearson correlation coefficient or Spearman rank correlation coefficient between different variables. Specifically, this can be implemented using the correlation analysis module in statistical software. Correlation coefficient analysis is used to screen samples with a significant statistical correlation to flood control benefits, eliminating interference from irrelevant variables.

[0056] Grey relational analysis can be understood as measuring the influence weight of different samples on the flood control benefit target by calculating the grey relational degree between each scheduling sample sequence and the ideal flood control benefit sequence. Specifically, it can be implemented using the relational degree calculation model in grey system theory. Correlation coefficient analysis is suitable for handling nonlinear relationships or small sample data, identifying key samples through geometric curve similarity.

[0057] In some embodiments, machine learning methods are used to fit the mapping relationships between samples to obtain the flood control benefit function of the reservoir for downstream flood control objectives under the scenario of random water inflow. This includes: fitting the mapping relationships between the features of each sample using a multiple linear regression model, a regression tree model, a boosting tree model, and a support vector machine model, respectively, to obtain the flood control benefit function corresponding to the multiple linear regression model, the regression tree model, the boosting tree model, and the support vector machine model; and selecting the flood control benefit function of the reservoir for downstream flood control objectives under the scenario of random water inflow from the flood control benefit functions corresponding to the multiple linear regression model, the regression tree model, the boosting tree model, and the support vector machine model.

[0058] In this embodiment, the multiple linear regression model can be understood as a statistical method that describes the relationship between independent and dependent variables through linear equations. Specifically, the least squares method can be used for parameter estimation to capture the linear correlation between variables in the flood control sample.

[0059] A regression tree model can be understood as a nonlinear regression model built based on the principle of decision trees. Specifically, it can be implemented using a recursive partitioning algorithm to handle the piecewise nonlinear relationship between flood characteristics and flood control benefits. The regression tree model can be a Bayesian decision tree model.

[0060] Boosting tree models can be understood as machine learning models that enhance predictive capabilities by integrating multiple regression trees. Specifically, they can be iteratively optimized using gradient boosting algorithms to improve the stability and generalization ability of flood control benefit predictions. One type of boosting tree model is the Bayesian boosting tree model.

[0061] Support Vector Machine (SVM) models can be understood as supervised learning methods that search for the optimal hyperplane in a high-dimensional space. Specifically, kernel functions can be used to handle nonlinear mappings, and this approach can be used to solve the generalization problem under high-dimensional features of flood data. The SVM model can be a Bayesian-SVM model.

[0062] Specifically, in flood control sample data, the multiple linear regression model analyzes the proportional relationship between upstream reservoir capacity input and downstream peak reduction flow by establishing a linear equation. For example, it can effectively represent the relationship when peak flow lag is linearly correlated with reservoir capacity demand. The regression tree model, by recursively partitioning the data space, can identify the optimal flood control benefit calculation method corresponding to different flood characteristic intervals. For example, it automatically switches to a piecewise calculation mode when the peak flow exceeds a threshold. The boosting tree model gradually corrects the prediction error of the preceding model by stacking multiple weak learners. For example, it improves prediction accuracy through multiple iterations when the peak flow combination features are complex. The support vector machine model uses kernel functions to map low-dimensional nonlinear data to a high-dimensional space. For example, it uses radial basis kernel functions to handle the nonlinear coupling relationship between peak flow and reservoir capacity. Through parallel computation of the four types of models, the linear, nonlinear, high-dimensional, and combined feature patterns that may exist in the flood control sample can be covered. Finally, the optimal model is selected as the target flood control benefit function based on the cross-validation error or coefficient of determination index.

[0063] In this application, a multi-model collaborative modeling and optimization mechanism is employed. This retains the interpretability advantages of linear models while incorporating the ability of machine learning models to represent complex relationships, effectively enhancing the adaptability of the flood control benefit function to different flood scenarios. It accurately establishes a flood control benefit function adapted to stochastic flood conditions, solving the problem of insufficient adaptability of reservoir capacity allocation schemes caused by the single model in traditional methods. Multi-model collaborative modeling covers various characteristic patterns in flood control, and the selection mechanism ensures the selection of the optimal model, enabling the flood control reservoir capacity allocation scheme to dynamically respond to different inflow scenarios, thus improving the scientific rigor and reliability of reservoir group flood control scheduling.

[0064] S140. Using the flood control benefit function as a constraint condition for the real-time optimization allocation model, a real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the goal of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed to obtain the reservoir group's capacity optimization allocation scheme according to different peak flow or water level targets, and the reliability of the calculation scheme is verified and evaluated.

[0065] In this embodiment, this application can establish a real-time optimization allocation model for flood control capacity of a reservoir group coupled with the flood control benefit function of the reservoir, design a model traversal solution strategy, and verify and evaluate the reliability of the model results. Then, the error distribution law can be obtained. If the error distribution is always positively or negatively skewed, the error correction coefficient can be increased to correct the results.

[0066] Specifically, this application achieves dynamic and precise allocation of flood control storage capacity, automatically adjusting the capacity allocation ratio of each reservoir based on real-time flood evolution. The established flood control benefit function transforms complex hydraulic response relationships into computable mathematical constraints, significantly improving the solution efficiency of the optimization model. The multi-reservoir collaborative optimization mechanism overcomes the overall benefit loss problem caused by local optimization in traditional methods, maximizing the utilization rate of flood control resources of the reservoir group while ensuring downstream flood control safety.

[0067] In some embodiments, the objective function of the real-time optimization allocation model includes maximizing the remaining flood control capacity of the reservoir group; the constraints of the real-time optimization allocation model also include at least one of the following: downstream flood control control point peak reduction constraint, downstream flood control control point water level constraint, and upstream reservoir capacity constraint.

[0068] In this embodiment, the downstream flood control control points can be jointly controlled by a group of reservoirs. The objectives of reducing water level and flow at the downstream flood control control points are transformed into model boundary constraints. A reserved capacity optimization allocation model is established with the objective of maximizing the remaining flood control capacity, thus transforming the multi-objective problem into a single-objective optimization problem. The maximum remaining flood control capacity of the reservoir group can be:

[0069] The target flood control benefit function can be:

[0070] The peak reduction constraint at downstream flood control control points can be:

[0071] The water level constraint at the downstream flood control control point can be:

[0072] The upstream reservoir capacity constraint can be:

[0073] Where, in the formula, and These represent the nonlinear mapping relationships between the reservoir capacity input to the k-th flood control station and the peak reduction at the downstream control station, respectively, and between the reservoir capacity input and the water level reduction at the downstream control station. The predictive factors considered by machine learning methods are upstream inflow, initial flow at downstream control stations, evolution lag time, and remaining storage capacity percentage, etc. and These are the peak reduction amount and water level reduction value at the control station, respectively. , , These are the upstream reservoir's capacity (i.e., the capacity that needs to be reserved), the remaining flood control capacity, and the total flood control capacity; and These are the peak flow rate of the control station before dispatch and the safe flow rate of the control station, respectively. and These are the peak flood level at the control station before dispatch and the safe water level at the control station, respectively.

[0074] Specifically, in solving the real-time optimization allocation model, the main parameter that needs to be optimized in this application is the peak reduction flow undertaken by each reservoir for the downstream flood control points. or water level reduction If the model involves a small number of upstream reservoirs and downstream flood control points (generally N and M are less than 2), a grid search method can be used to solve the problem. That is, when the model involves two reservoirs (Reservoir A and Reservoir B) and one flood control station C, the peak flow reduction that Reservoir A needs to provide to the downstream flood control station can be discretized first. The corresponding reservoir B needs to reduce the peak flow at the downstream flood control point by the following amount: ,according to and Calculate the required storage capacity for each reservoir group, then calculate the sum of the remaining storage capacity. By repeating this process, find a solution that maximizes the remaining storage capacity of the reservoir group.

[0075] When there are many parameters to optimize, intelligent algorithms can be used to optimize the peak flow reduction of each reservoir to downstream flood control points. The solution is obtained by iteratively finding the optimal allocation scheme for the remaining flood control capacity of the reservoir group that satisfies the constraints.

[0076] In some embodiments, the reliability verification and evaluation of the calculation scheme includes: using dynamic programming or linear programming methods to establish a deterministic optimization scheduling model with the objective of maximizing the remaining flood control capacity, in order to optimize the reservoir discharge process and obtain the optimal reservoir capacity optimization allocation result; comparing the optimal reservoir capacity optimization allocation result with the reservoir capacity optimization allocation scheme to obtain the error distribution pattern between the two, and performing reliability verification and evaluation based on the error distribution pattern.

[0077] This application also requires verification of the reservoir capacity optimization allocation scheme. Specifically, dynamic programming, linear programming, and other methods can be used to establish a deterministic optimization scheduling model with the objective of maximizing remaining flood control capacity. This optimizes the reservoir discharge process, yields the optimal reservoir capacity allocation result, and compares this result with the allocation result calculated by the real-time optimization allocation model to calculate the error and provide the error distribution pattern. If the error distribution is consistently positively or negatively skewed, an error correction coefficient can be added to correct the result. Dynamic programming can be understood as breaking down a large problem into a series of interconnected smaller problems, solving them sequentially from back to front (or from front to back), and remembering intermediate results to avoid redundant calculations. Linear programming can be understood as listing the problem as a set of linear equations / inequalities and finding the vertex in a high-dimensional polyhedron that best achieves a certain linear objective.

[0078] The technical solution of this application will be described in detail below with reference to a specific embodiment.

[0079] The Xiluodu, Xiangjiaba, and Pubugou reservoirs, a combined reservoir group, were selected as the research object to verify the rationality of the model. Xiluodu and Xiangjiaba were treated as virtual reservoirs connected in parallel with Pubugou Reservoir to jointly ensure the flood control safety of the downstream Lizhuang station. Based on historical and simulated flood sequences, a real-time optimization allocation method for the flood control reserve capacity of the reservoir group considering random inflows was implemented. The specific implementation steps are as follows: S1. Based on the distances of the Xi, Xiang, and Pubugou reservoirs from the Lizhuang station, and considering the differences in their downstream peak-shaving compensation capabilities, different dispatching modes were formulated. Since the Xi and Xiang reservoirs are closer to Lizhuang and provide better compensation, the peak-shaving compensation dispatching mode was adopted as the framework. However, the Pubugou reservoir is farther from Lizhuang, so an equal-capacity interception method was used to intercept the excess flood peak and volume at Lizhuang. Using the predicted flow rate of Lizhuang station for the next day as available information, and employing the framework described in Table 1, optimized dispatching models were established for the Xi, Xiang, and Pubugou reservoirs respectively. Furthermore, 500 simulations of a 50-year return period multi-station overall design flood process controlled by Lizhuang were performed. The simulation data were input into the optimized dispatching model to obtain a flood control dispatching sample set, and the reservoir flood control benefit curves were extracted for correction.

[0080] S2. The scheduling results were analyzed using grey relational analysis to extract key flood features with high correlation to reservoir capacity allocation. Seven flood features were selected as key feature vectors to identify different types of floods: upstream reservoir forecast peak flow, downstream control point forecast peak flow, peak flow delay, and the percentage of the maximum 1-day, 3-day, 7-day, and 15-day flood volume downstream. Training samples were constructed by combining the different peak reduction amounts for each flood with the corresponding upstream reservoir capacity required. Four methods—multiple linear regression, Bayesian support vector machine, Bayesian decision tree, and Bayesian boosting tree—were used to fit the reservoir flood control benefit response relationship. Comparative analysis showed that the Bayesian boosting tree coupled model had the best fitting effect.

[0081] Table 3. Grey Relational Analysis Results of Predictive Factors Selected for Different Reservoirs

[0082] Table 4. Grey Relational Analysis Results of Predictive Factors Selected for Different Reservoirs

[0083] S3. Establish a reservoir capacity optimization allocation model based on maximizing remaining flood control capacity. Select seven typical 15-day flood events from 1961, 1966, 1981, 1991, 1998, 2010, and 2012, with Lizhuang as the control point, scaled up to a 50-year return period. Input the data into the model to obtain the optimal flood control capacity allocation scheme. Compare and analyze the optimization results of the deterministic joint optimization model of the reservoir group with the theoretical optimal capacity allocation scheme. The results are shown in Table 5 below.

[0084] Table 5 Results of reservoir capacity optimization allocation under typical annual conditions considering random water inflow.

[0085] As shown in Table 5, the reservoir capacity allocation results obtained based on the optimal flood control benefit allocation model are similar to those obtained from joint optimization scheduling. Under typical annual water conditions, the error between the two is small, indicating high model reliability. The proposed model not only considers seven typical design flood processes but also, through stochastic simulation techniques, further considers typical design flood processes with different regional compositions, effectively overcoming the problem of insufficient diversity in design floods.

[0086] In the real-time optimization allocation method for flood control reserved capacity of reservoir groups considering random inflow provided in this application embodiment, a first sample set of reservoir optimization scheduling for random simulated floods is generated based on a pre-set reservoir flood control rule optimization scheduling model. Around the first scheduling sample set, a cluster of extracted flood control scheduling benefit curves is defined and modified to obtain a second scheduling sample set for random simulated floods. Key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input are identified in the second scheduling sample set. Machine learning methods are used to fit the mapping relationship between each sample feature to obtain the flood control benefit function of the reservoir for downstream flood control control objectives under the scenario of random inflow. The objective flood control benefit function is used as a constraint condition for the real-time optimization allocation model. A real-time optimization allocation model for reserved flood control capacity of the reservoir group is constructed with the objective of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed, and then the reservoir group capacity optimization allocation scheme is obtained according to different peak reduction flow / water level objectives. The reliability of the calculation scheme is verified and evaluated. The proposed method solves the problem that the rationality of the flood control capacity optimization allocation scheme results is low due to the single design flood process.

[0087] This application also provides a real-time optimized allocation device 200 for flood control reserved capacity of a reservoir group that takes into account random water inflow. This device is used to execute any embodiment of the aforementioned real-time optimized allocation method for flood control reserved capacity of a reservoir group that takes into account random water inflow.

[0088] Specifically, please refer to Figure 3 , Figure 3 This is a schematic block diagram of a real-time optimized allocation device 200 for flood control reserve capacity of a reservoir group that takes into account random water inflow, provided in an embodiment of this application.

[0089] like Figure 3 As shown, the real-time optimization allocation device 200 for flood control reserve capacity of reservoir groups considering random water inflow provided in this application includes: a first generation unit 210, a correction unit 220, a fitting unit 230, and a second generation unit 240.

[0090] The first generation unit 210 is used to generate a first scheduling sample set for reservoir optimization scheduling oriented to random simulated floods based on a pre-set reservoir flood control rule optimization scheduling model. The correction unit 220 is used to define and extract a cluster of flood control scheduling benefit curves around the first scheduling sample set, and correct the cluster of flood control scheduling benefit curves to obtain a second scheduling sample set for random simulated floods. The fitting unit 230 is used to identify the sample features in the second scheduling sample set, and use machine learning methods to fit the mapping relationship between each sample feature to obtain the flood control benefit function of the reservoir for downstream flood control control objectives under random inflow scenarios. The sample features include a set of key flood feature factors, downstream peak reduction amount, and upstream reservoir input capacity. The second generation unit 240 is used to use the flood control benefit function as a constraint condition for the real-time optimization allocation model, construct a real-time optimization allocation model for the reserved flood control capacity of the reservoir group with the objective of maximizing the remaining flood control capacity of the reservoir group, design a model traversal solution strategy to obtain the reservoir group capacity optimization allocation scheme according to different peak reduction flow or water level targets, and verify and evaluate the reliability of the calculation scheme.

[0091] The real-time optimization allocation device 200 for reservoir group flood control reserved capacity considering random inflow provided in this application embodiment can generate a first sample set of reservoir optimization scheduling for random simulated floods based on a pre-set reservoir flood control rule optimization scheduling model. Around the first scheduling sample set, a cluster of extracted flood control scheduling benefit curves is defined and corrected to obtain a second scheduling sample set for random simulated floods. Key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input in the second scheduling sample set are identified. Machine learning methods are used to fit the mapping relationship between each sample feature to obtain the flood control benefit function of the reservoir for downstream flood control control objectives under the random inflow scenario. The objective flood control benefit function is used as a constraint condition for the real-time optimization allocation model. A real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the objective of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed, and then the reservoir group capacity optimization allocation scheme is obtained according to different peak reduction flow / water level objectives. The reliability of the calculation scheme is verified and evaluated. The proposed method solves the problem that the rationality of the flood control capacity optimization allocation scheme results is low due to the single design flood process.

[0092] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the real-time optimization allocation device 200 for flood control reserve capacity of reservoir groups that takes into account random water inflow and each unit can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0093] The aforementioned real-time optimization allocation device 200 for flood control reserve capacity of a reservoir group that considers random water inflow can be implemented as a computer program, which can, for example... Figure 4It runs on the electronic device shown.

[0094] Please see Figure 4 , Figure 4 This is a schematic block diagram of the electronic device 300 provided in the embodiments of this application.

[0095] See Figure 4 The electronic device 300 includes a processor 302, a memory, and a network interface 305 connected via a system bus 301. The memory may include a storage medium 303 and internal memory 304.

[0096] The storage medium 303 can store the operating system 3031 and the computer program 3032. When the computer program 3032 is executed, it enables the processor 302 to execute a real-time optimized allocation method for the flood control reserved capacity of the reservoir group, taking into account random water inflows.

[0097] The processor 302 provides computing and control capabilities to support the operation of the entire device 300.

[0098] The internal memory 304 provides an environment for the operation of the computer program 3032 in the non-volatile storage medium 303. When the computer program 3032 is executed by the processor 302, the processor 302 can execute a real-time optimized allocation method for the flood control reserved capacity of the reservoir group that takes into account random water inflow.

[0099] The network interface 305 is used for network communication, such as providing data transmission. Those skilled in the art will understand that... Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the device 300 to which the present application is applied. The specific device 300 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0100] The processor 302 is used to run the computer program 3032 stored in the memory to achieve the following functions: Based on a pre-set reservoir flood control rule optimization scheduling model, a first sample set of reservoir optimization scheduling for random simulated floods is generated; around the first scheduling sample set, a cluster of extracted flood control scheduling benefit curves is defined and corrected to obtain a second scheduling sample set for random simulated floods; key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input in the second scheduling sample set are identified, and machine learning methods are used to fit the mapping relationship between each sample feature to obtain the flood control benefit function of the reservoir for downstream flood control control objectives under random inflow scenarios; the objective flood control benefit function is used as a constraint condition for the real-time optimization allocation model, and a real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the objective of maximizing the remaining flood control capacity of the reservoir group; a model traversal solution strategy is designed, and then the reservoir group capacity optimization allocation scheme is obtained according to different peak reduction flow / water level objectives, and the reliability of the calculation scheme is verified and evaluated; the proposed method solves the problem that the rationality of the flood control capacity optimization allocation scheme is low due to the single design flood process.

[0101] Those skilled in the art will understand that Figure 4 The embodiments of device 300 shown do not constitute a limitation on the specific configuration of device 300. In other embodiments, device 300 may include more or fewer components than shown, or combine certain components, or have different component arrangements. For example, in some embodiments, device 300 may include only a memory and processor 302. In such embodiments, the structure and function of the memory and processor 302 are similar to those shown. Figure 4 The embodiments shown are consistent and will not be repeated here.

[0102] It should be understood that in the embodiments of this application, the processor 302 may be a central processing unit (CPU), or it may be other general-purpose processors 302, digital signal processors 302 (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor 302 may be a microprocessor 302, or it may be any conventional processor 302, etc.

[0103] According to one aspect of this application, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. The processor of the electronic device reads the computer instructions from a computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the following steps: Based on a pre-set reservoir flood control rule optimization scheduling model, a first sample set of reservoir optimization scheduling for random simulated floods is generated; around the first scheduling sample set, a cluster of extracted flood control scheduling benefit curves is defined and corrected to obtain a second scheduling sample set for random simulated floods; key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input are identified in the second scheduling sample set, and machine learning methods are used to fit the mapping relationship between each sample feature to obtain the flood control benefit function of the reservoir for downstream flood control control objectives under random inflow scenarios; the objective flood control benefit function is used as a constraint condition for the real-time optimization allocation model, and a real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the objective of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed, and then the reservoir group capacity optimization allocation scheme is obtained according to different peak reduction flow / water level objectives, and the reliability of the calculation scheme is verified and evaluated; the proposed method solves the problem that the rationality of the flood control capacity optimization allocation scheme is low due to the single design flood process.

[0104] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0105] In another embodiment of this application, a computer storage medium is provided. This storage medium may be a non-volatile computer-readable storage medium or a volatile storage medium. The storage medium stores a computer program 3032, which, when executed by the processor 302, performs the following steps: Based on a pre-set reservoir flood control rule optimization scheduling model, a first sample set of reservoir optimization scheduling for random simulated floods is generated; around the first scheduling sample set, a cluster of extracted flood control scheduling benefit curves is defined and corrected to obtain a second scheduling sample set for random simulated floods; key flood characteristic factors, downstream peak reduction, and upstream reservoir capacity input are identified in the second scheduling sample set, and machine learning methods are used to fit the mapping relationship between each sample feature to obtain the flood control benefit function of the reservoir for downstream flood control control objectives under random inflow scenarios; the objective flood control benefit function is used as a constraint condition for the real-time optimization allocation model, and a real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the objective of maximizing the remaining flood control capacity of the reservoir group; a model traversal solution strategy is designed, and then the reservoir group capacity optimization allocation scheme is obtained according to different peak reduction flow / water level objectives, and the reliability of the calculation scheme is verified and evaluated; the proposed method solves the problem that the rationality of the flood control capacity optimization allocation scheme results is low due to the single design flood process.

[0106] The storage medium can be any computer-readable storage medium that can store program code, such as a USB flash drive, external hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0107] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0108] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0109] The steps in the methods of this application embodiment can be adjusted, merged, or deleted according to actual needs. The units in the apparatus of this application embodiment can be merged, divided, or deleted according to actual needs. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0110] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods provided in the various embodiments of this application.

[0111] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflow, characterized in that, include: Based on the pre-set reservoir flood control rule optimization scheduling model, the first scheduling sample set of reservoir optimization scheduling for random simulated floods is generated. Based on the first scheduling sample set, a cluster of flood control scheduling benefit curves is defined and extracted, and the cluster of flood control scheduling benefit curves is modified to obtain a second scheduling sample set for random simulated floods; The sample features in the second scheduling sample set are identified, and the mapping relationship between each sample feature is fitted using machine learning methods to obtain the flood control benefit function of the reservoir for the downstream flood control target under the scenario of random water inflow; the sample features include the key flood feature factor set, the downstream peak reduction amount and the upstream reservoir input capacity; Using the flood control benefit function as a constraint condition for the real-time optimization allocation model, a real-time optimization allocation model for the reserved flood control capacity of the reservoir group is constructed with the goal of maximizing the remaining flood control capacity of the reservoir group. A model traversal solution strategy is designed to obtain the reservoir group's capacity optimization allocation scheme according to different peak flow or water level targets, and the reliability of the calculation scheme is verified and evaluated.

2. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows, as described in claim 1, is characterized in that... Before generating the first scheduling sample set for reservoir optimization scheduling oriented towards stochastic simulated floods using the preset reservoir flood control rule optimization scheduling model, the following steps are also included: Determine the flood control scheduling rules for upstream reservoirs; According to the scheduling rules, the scheduling discrimination index of the upstream reservoir for the downstream flood control point, the corresponding scheduling method of the upstream reservoir, and the classification method of the scheduling discrimination index are determined. The scheduling discrimination index includes at least one of the following: real-time inflow index of the reservoir, forecasted rainfall index, real-time and forecasted water level and flow index of the downstream flood control point, and current water level index of the reservoir. The scheduling method includes equal storage and compensation scheduling. Based on the scheduling discrimination index, the hierarchical method, and the scheduling method, the reservoir flood control rule optimization scheduling model is constructed.

3. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows, as described in claim 2, is characterized in that... The step of constructing the reservoir flood control rule optimization scheduling model based on the scheduling discrimination index, the hierarchical method, and the scheduling method includes: The scheduling discrimination index is classified into different levels, and the reservoir scheduling mode under different classification discrimination indexes is parameterized and modeled to determine the flood control rule optimization scheduling objective. The optimization algorithm is then used to solve the classification discrimination index and scheduling mode modeling parameters.

4. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows, as described in claim 1, is characterized in that... The objective function of the reservoir flood control rule optimization scheduling model includes the cumulative remaining flood control capacity of the upstream reservoir; The constraints of the reservoir flood control rule optimization scheduling model include at least one of the following: water level and flow constraints at the downstream flood control control point, water balance constraints at the inflow and outflow of the upstream reservoir, and maximum and minimum water level and flow constraints.

5. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows, as described in any one of claims 1-4, is characterized in that... The optimized scheduling model based on pre-set reservoir flood control rules generates a first scheduling sample set of random simulated floods, including: A pre-set multi-station flood simulation model is used to perform stochastic simulation of the flood process in order to obtain key characteristic information of the simulated flood; the key characteristic information includes the flood peak of the upstream reservoir, the flood peak of the downstream flood control point, and the flood peak delay time. Based on the key feature information, the upstream reservoir is optimized and scheduled using the reservoir flood control rule optimization scheduling model to obtain the upstream reservoir's capacity input, the peak flow reduction at the downstream flood control control point, and the maximum water level reduction. The first scheduling sample set is obtained by combining the key feature information, the upstream reservoir's capacity, the peak flow reduction and the highest water level reduction at the downstream flood control point.

6. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows according to any one of claims 1-4, characterized in that, The process of identifying sample features in the second scheduling sample set and fitting the mapping relationship between each sample feature using machine learning methods to obtain a flood control benefit function for the reservoir's downstream flood control objectives under a random inflow scenario includes: The second scheduling sample set is analyzed and screened using correlation coefficient analysis or grey relational analysis to construct sample set data; Machine learning methods are used to fit the mapping relationship between the features of each sample in the sample set data to obtain the flood control benefit function of the reservoir for the downstream flood control target under the scenario of random water inflow.

7. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows, as described in claim 6, is characterized in that... The step of using correlation coefficient analysis or grey relational analysis to analyze and filter the second scheduling sample set to construct sample set data includes: The second scheduling sample set was analyzed and screened using correlation coefficient analysis or grey relational analysis to identify and extract key flood characteristics related to the reservoir's capacity input. Establish set-pair relationships between key flood characteristics, downstream peak reduction, and upstream reservoir capacity to construct the sample set data.

8. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows according to any one of claims 1-4, characterized in that, The method employs machine learning to fit the mapping relationship between the features of each sample, resulting in a flood control benefit function for the reservoir considering random inflow scenarios and targeting downstream flood control objectives. This function includes: The mapping relationship between the features of each sample is fitted by a multiple linear regression model, a regression tree model, a boosting tree model, and a support vector machine model, respectively, to obtain the flood control benefit function corresponding to the multiple linear regression model, the flood control benefit function corresponding to the regression tree model, the flood control benefit function corresponding to the boosting tree model, and the flood control benefit function corresponding to the support vector machine model. From the flood control benefit functions corresponding to the multiple linear regression model, the regression tree model, the boosting tree model, and the support vector machine model, the flood control benefit function of the reservoir for downstream flood control objectives under the scenario of random water inflow is selected.

9. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows according to any one of claims 1-4, characterized in that, The objective function of the real-time optimization allocation model includes maximizing the remaining flood control capacity of the reservoir group; The constraints of the real-time optimization allocation model also include at least one of the following: downstream flood control point peak reduction constraint, downstream flood control point water level constraint, and upstream reservoir capacity constraint.

10. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows according to any one of claims 1-4, characterized in that, The verification and evaluation of the reliability of the computation scheme includes: A deterministic optimization scheduling model with the objective of maximizing the remaining flood control capacity is established using dynamic programming or linear programming methods to optimize the reservoir discharge process and obtain the optimal reservoir capacity allocation result. The optimal storage capacity allocation result is compared with the storage capacity allocation scheme to obtain the error distribution pattern between the two, and the reliability is verified and evaluated based on the error distribution pattern.

11. The real-time optimization allocation method for flood control reserve capacity of a reservoir group considering random inflows according to any one of claims 1-4, characterized in that, The process of correcting the flood control scheduling benefit curve cluster yields a second scheduling sample set for stochastic simulated floods, including: Develop a correction strategy for the aforementioned flood control scheduling benefit curve cluster; Based on the aforementioned correction strategy, discrete points are inserted into the cluster of flood control scheduling benefit curves to correct the cluster of flood control scheduling benefit curves, thereby obtaining a second scheduling sample set for random simulated floods.

Images