A river network water conservancy dispatching method and device, computer equipment and storage medium

Abstract

The application provides a river network water engineering scheduling method and device, computer equipment and storage medium, wherein river network basic data is acquired to determine key control points and safety thresholds thereof, and a joint probability distribution model of uncertainty parameters is established; a random scene set is generated based on the model to obtain response results of the key control points through hydrodynamic simulation; sample frequencies of the key control points satisfying the safety thresholds are counted, and a lower confidence bound thereof under a given confidence level is calculated, so that the condition that the lower confidence bound is not lower than a target safety level is used as a satisfaction condition of a probabilistic safety constraint; when the lower confidence bound is lower than the target safety level or a confidence interval width exceeds a preset threshold, random scenes are added and repeated until the constraint is satisfied or a sample upper limit is reached; and an optimization model with water engineering operation parameters as decision variables is solved to generate a scheduling scheme and is executed under the satisfaction of the constraint. By using the above method, the probabilistic safety guarantee reliability, calculation efficiency and real-time closed-loop adaptability of river network water engineering scheduling can be improved.

Description

Technical Field

[0001] This application relates to the field of water conservancy project scheduling technology, and more specifically, to a method, apparatus, computer equipment, and storage medium for scheduling river network water projects. Background Technology

[0002] The core requirement in the field of water conservancy project scheduling is to ensure the safe and stable operation of river network systems and water conservancy projects, and to balance comprehensive objectives such as flood control, water supply, power generation, and ecological water demand. River network hydrodynamic simulation is the foundation of scheduling decisions, and it needs to reflect the changing patterns of key hydrological indicators such as water level and flow. Water project scheduling, on the other hand, needs to formulate control strategies for facilities such as gates, pumping stations, and reservoirs based on simulation results. The coordinated cooperation between the two is the core technical direction in this field.

[0003] In existing technologies, river network hydrodynamic simulation mainly relies on one-dimensional or two-dimensional hydrodynamic models (such as the Saint-Venant equations). By inputting deterministic data such as measured or predicted rainfall, water level, and flow, the dynamic processes of water level and flow within the river network are calculated. Based on this, water project scheduling often adopts regular curves, empirical scheduling methods, or simple deterministic optimization techniques (such as linear programming). Control schemes such as gate opening and pump station start-up and shutdown are generated according to preset fixed thresholds or empirical formulas to meet preset safety or benefit objectives.

[0004] The aforementioned existing technologies have significant drawbacks: insufficient ability to cope with uncertainties. River network operations are actually affected by various uncertainties such as fluctuations in water inflow, changes in river channel roughness, and deviations in engineering execution. However, traditional technologies often treat input data as deterministic values ​​or rely solely on subjectively setting fixed safety margins to address uncertainties. This fails to quantify the probability distribution of "over-limit risks" and makes it difficult to scientifically determine safety levels. Consequently, the generated scheduling schemes lack robustness, and when actual operating conditions deviate from preset conditions, safety hazards or insufficient resource utilization are likely to occur. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide a method, apparatus, computer equipment and storage medium for scheduling river network water projects, which can improve the reliability of probabilistic safety assurance, computational efficiency and real-time closed-loop adaptability of river network water project scheduling.

[0006] In a first aspect, embodiments of this application provide a method for scheduling river network water conservancy projects, the method comprising:

[0007] Acquire basic river network data, determine key control points and their safety thresholds based on the basic river network data, and establish a joint probability distribution model of uncertain input parameters;

[0008] Based on the joint probability distribution model, a random scenario set is generated, and the response results of the key control points under each scenario are obtained through hydrodynamic simulation.

[0009] Based on the response results, the frequency of samples that the key control points meet the security threshold is statistically analyzed, and the lower bound of the confidence level is calculated based on the sample frequency. The lower bound of the confidence level is not lower than the preset target security level as the condition for satisfying the probabilistic security constraint.

[0010] When the lower confidence bound is lower than the preset target safety level or the confidence interval width of the sample frequency exceeds the preset threshold, additional random scenarios are generated and the hydrodynamic simulation and statistical verification steps are repeated until the probabilistic safety constraint is met or the preset sample upper limit is reached.

[0011] Under the premise of satisfying the probabilistic safety constraints, an optimization model with water project operation parameters as decision variables is solved to generate a scheduling scheme and execute the scheduling scheme.

[0012] Optionally, determining key control points and their safety thresholds based on the river network basic data includes:

[0013] The river network basic data is preprocessed to obtain preprocessed data;

[0014] Based on the preprocessed data, key control points are selected from river cross-sections, hydraulic structures, hydrological stations, and important protected objects;

[0015] Set a corresponding safety threshold for each of the aforementioned critical control points.

[0016] Optionally, the step of generating a random scenario set based on the joint probability distribution model and obtaining the response results of the key control points under each scenario through hydrodynamic simulation includes:

[0017] The joint probability distribution model is sampled using quasi-Monte Carlo low-discrepancy sequences to obtain multiple random scenarios, forming an initial random scenario set;

[0018] For each random scene in the initial random scene set, the sampled random scene parameters are mapped to the boundary conditions and model parameters of the hydrodynamic model to obtain the mapped input conditions.

[0019] Based on the mapped input conditions, the river network hydrodynamic model is run to obtain the water level sequence and flow sequence of each key control point during the simulation period.

[0020] The response results of each key control point are extracted from the water level sequence and flow rate sequence.

[0021] Optionally, the step of statistically analyzing the sample frequency of the key control points satisfying the security threshold based on the response results includes:

[0022] For each key control point, determine whether its response result in each random scenario meets the security threshold, and obtain the satisfaction status in each scenario.

[0023] Count the number of scenarios where the satisfaction state is "yes" to obtain the satisfaction count;

[0024] The ratio of the number of times the condition is satisfied to the total number of scenarios is used as the sample frequency.

[0025] Optionally, calculating the lower confidence bound of the sample at a given confidence level based on the sample frequency includes:

[0026] Obtain the sample frequency and the corresponding total number of random scenarios;

[0027] A method for constructing precise confidence intervals using binomial distribution proportions is employed, based on the sample frequency, the total number of random scenarios, and the given confidence level, to calculate the confidence intervals.

[0028] The lower limit of the confidence interval is used as the lower confidence bound.

[0029] Optionally, the step of solving the optimization model with water project operation parameters as decision variables to generate a scheduling scheme, under the premise of satisfying the probabilistic safety constraints, includes:

[0030] An objective function is constructed by using at least one of the following as decision variables: gate opening sequence, pump station start-up and shutdown scheme, and reservoir inflow and outflow.

[0031] The probabilistic safety constraints are used as constraints on the optimization model.

[0032] Based on the objective function and the constraints, the optimization model is solved using mathematical programming or intelligent optimization algorithms to obtain the optimized values ​​of the decision variables.

[0033] Based on the optimized decision variable values, an executable water project scheduling scheme is generated.

[0034] Optionally, the method further includes at least one of the following:

[0035] The scheduling scheme is executed in a rolling time domain to obtain real-time monitoring data during the execution process;

[0036] The joint probability distribution model is updated online based on the real-time monitoring data to obtain the updated joint probability distribution model.

[0037] Based on the updated joint probability distribution model, the step of generating the random scene set is returned, forming a closed-loop operation;

[0038] Record the risk level, sample size, statistical confidence interval, and execution effect of each scheduling decision to form a traceable scheduling decision log;

[0039] When real-time monitoring data is missing or delayed, online inference is performed based on the established joint probability distribution model to obtain inferred data, and the continuity of scheduling operation is maintained based on the inferred data.

[0040] Secondly, embodiments of this application provide a river network water conservancy project scheduling device, the device comprising:

[0041] The model building module is used to acquire basic river network data, determine key control points and their safety thresholds based on the basic river network data, and establish a joint probability distribution model of uncertain input parameters.

[0042] The response result generation module is used to generate a random scenario set based on the joint probability distribution model, and obtain the response results of the key control points under each scenario through hydrodynamic simulation.

[0043] The condition determination module is used to statistically analyze the sample frequency of the key control points satisfying the security threshold based on the response results, and calculate the lower confidence bound of the sample frequency at a given confidence level, with the lower confidence bound not being lower than the preset target security level as the condition for satisfying the probabilistic security constraint.

[0044] The statistical verification module is used to generate additional random scenarios and repeat the hydrodynamic simulation and statistical verification steps when the confidence lower bound is lower than the preset target safety level or the confidence interval width of the sample frequency exceeds the preset threshold, until the probability safety constraint is met or the preset sample upper limit is reached.

[0045] The scheduling scheme execution module is used to solve the optimization model with water project operation parameters as decision variables, generate a scheduling scheme, and execute the scheduling scheme, under the premise of satisfying the probabilistic safety constraints.

[0046] Optionally, determining key control points and their safety thresholds based on the river network basic data includes:

[0047] The river network basic data is preprocessed to obtain preprocessed data;

[0048] Based on the preprocessed data, key control points are selected from river cross-sections, hydraulic structures, hydrological stations, and important protected objects;

[0049] Set a corresponding safety threshold for each of the aforementioned critical control points.

[0050] Optionally, the step of generating a random scenario set based on the joint probability distribution model and obtaining the response results of the key control points under each scenario through hydrodynamic simulation includes:

[0051] The joint probability distribution model is sampled using quasi-Monte Carlo low-discrepancy sequences to obtain multiple random scenarios, forming an initial random scenario set;

[0052] For each random scene in the initial random scene set, the sampled random scene parameters are mapped to the boundary conditions and model parameters of the hydrodynamic model to obtain the mapped input conditions.

[0053] Based on the mapped input conditions, the river network hydrodynamic model is run to obtain the water level sequence and flow sequence of each key control point during the simulation period.

[0054] The response results of each key control point are extracted from the water level sequence and flow rate sequence.

[0055] Optionally, the step of statistically analyzing the sample frequency of the key control points satisfying the security threshold based on the response results includes:

[0056] For each key control point, determine whether its response result in each random scenario meets the security threshold, and obtain the satisfaction status in each scenario.

[0057] Count the number of scenarios where the satisfaction state is "yes" to obtain the satisfaction count;

[0058] The ratio of the number of times the condition is satisfied to the total number of scenarios is used as the sample frequency.

[0059] Optionally, calculating the lower confidence bound of the sample at a given confidence level based on the sample frequency includes:

[0060] Obtain the sample frequency and the corresponding total number of random scenarios;

[0061] A method for constructing precise confidence intervals using binomial distribution proportions is employed, based on the sample frequency, the total number of random scenarios, and the given confidence level, to calculate the confidence intervals.

[0062] The lower limit of the confidence interval is used as the lower confidence bound.

[0063] Optionally, the step of solving the optimization model with water project operation parameters as decision variables to generate a scheduling scheme, under the premise of satisfying the probabilistic safety constraints, includes:

[0064] An objective function is constructed by using at least one of the following as decision variables: gate opening sequence, pump station start-up and shutdown scheme, and reservoir inflow and outflow.

[0065] The probabilistic safety constraints are used as constraints on the optimization model.

[0066] Based on the objective function and the constraints, the optimization model is solved using mathematical programming or intelligent optimization algorithms to obtain the optimized values ​​of the decision variables.

[0067] Based on the optimized decision variable values, an executable water project scheduling scheme is generated.

[0068] Optionally, the apparatus further includes an execution monitoring module for performing at least one of the following methods:

[0069] The scheduling scheme is executed in a rolling time domain to obtain real-time monitoring data during the execution process;

[0070] The joint probability distribution model is updated online based on the real-time monitoring data to obtain the updated joint probability distribution model.

[0071] Based on the updated joint probability distribution model, the step of generating the random scene set is returned, forming a closed-loop operation;

[0072] Record the risk level, sample size, statistical confidence interval, and execution effect of each scheduling decision to form a traceable scheduling decision log;

[0073] When real-time monitoring data is missing or delayed, online inference is performed based on the established joint probability distribution model to obtain inferred data, and the continuity of scheduling operation is maintained based on the inferred data.

[0074] Thirdly, embodiments of this application provide a computer device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, the steps of the river network water engineering scheduling method described in any of the optional embodiments of the first aspect are performed.

[0075] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the river network water engineering scheduling method described in any of the optional embodiments of the first aspect.

[0076] The technical solution provided in this application includes, but is not limited to, the following beneficial effects:

[0077] The process involves acquiring basic river network data, determining key control points and their safety thresholds based on this data, and establishing a joint probability distribution model for uncertain input parameters. This step achieves standardized preprocessing of multi-source basic data, resolving the inconsistency between spatiotemporal and measurement standards in traditional data, and providing consistent data support. It also clarifies core key control points and quantifies safety thresholds, establishing unified risk assessment criteria to avoid the ambiguity of single-value constraints / empirical margins. Furthermore, it systematically characterizes the marginal distribution and correlation structure of uncertain parameters, filling the gap in traditional technologies' lack of systematic characterization of uncertainty.

[0078] A random scenario set is generated based on the joint probability distribution model, and the response results of the key control points in each scenario are obtained through hydrodynamic simulation. This step generates a random scenario set covering the uncertainty space, solving the problem of incomplete coverage in traditional fixed sample simulation. The water level, flow sequence and risk indicators of key control points in each scenario are obtained to avoid one-sided simulation data. Costs are controlled by optimizing calculation methods, alleviating the pain point of high cost in traditional hydrodynamic simulation.

[0079] Based on the response results, the frequency of samples where the key control points meet the safety threshold is statistically analyzed, and a lower confidence bound at a given confidence level is calculated based on the sample frequency. The lower confidence bound being no lower than the preset target safety level is used as the condition for satisfying the probabilistic safety constraint. Through this step, the safety assessment is transformed from qualitative to quantitative. The lower confidence bound is used for conservative verification to avoid risk misjudgment caused by traditional sample point estimation, and verifiable determination of probabilistic safety constraints under limited samples is achieved. The constraint satisfaction criteria are clearly defined, providing a clear and reliable safety basis for scheduling optimization.

[0080] When the lower confidence bound is lower than the preset target safety level or the confidence interval width of the sample frequency exceeds the preset threshold, additional random scenarios are generated and the hydrodynamic simulation and statistical verification steps are repeated until the probability safety constraint is met or the preset sample upper limit is reached. Through this step, an adaptive sample addition closed-loop mechanism is constructed, adding scenarios only when necessary to reduce invalid computational overhead; the sample size is dynamically adjusted to balance computational accuracy and solution timeliness, solving the problem of insufficient real-time performance in traditional fixed large-sample simulations.

[0081] Under the premise of satisfying the probabilistic safety constraints, an optimization model with water project operation parameters as decision variables is solved to generate a scheduling scheme and execute the scheduling scheme. Through this step, a multi-objective optimization model is constructed to ensure that the scheduling scheme takes into account both safety and comprehensive benefits, and avoids the imbalance between safety and benefits. Directly executable control commands are output to realize the integrated processing of engineering constraints and solve the problem of poor feasibility of traditional schemes. A rolling scheduling strategy is adopted to dynamically update the model and scheme, thereby improving the adaptability and robustness under extreme conditions.

[0082] The five inventions form a complete technical chain, which respectively solves the core problems in traditional technologies such as data disorder, insufficient uncertainty representation, misjudgment of risk assessment, excessive computing costs, disconnect between solutions and poor adaptability. Ultimately, it achieves the integrated technical effect of probabilistic safety verifiable, computing cost controllable and online closed-loop operability, which can improve the reliability of probabilistic safety assurance, computing efficiency and real-time closed-loop adaptability of river network hydrodynamic simulation and water project scheduling under uncertain conditions.

[0083] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0084] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0085] Figure 1 A flowchart of a river network water conservancy project scheduling method provided in Embodiment 1 of this application is shown;

[0086] Figure 2 A flowchart of a security threshold determination method provided in Embodiment 1 of this application is shown;

[0087] Figure 3 A flowchart of a response result generation method provided in Embodiment 1 of this application is shown;

[0088] Figure 4 A flowchart of a sample frequency generation method provided in Embodiment 1 of this application is shown;

[0089] Figure 5 A flowchart of a confidence lower bound determination method provided in Embodiment 1 of this application is shown;

[0090] Figure 6 A flowchart of a scheduling scheme generation method provided in Embodiment 1 of this application is shown;

[0091] Figure 7 This paper shows a schematic diagram of the structure of a river network water engineering scheduling device provided in Embodiment 2 of this application;

[0092] Figure 8 A schematic diagram of the structure of a computer device provided in Embodiment 3 of this application is shown. Detailed Implementation

[0093] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, 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, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0094] Example 1

[0095] To facilitate understanding of this application, the following is combined with... Figure 1 The flowchart illustrating the river network water engineering scheduling method provided in Embodiment 1 of this application will be used to describe Embodiment 1 of this application in detail.

[0096] See Figure 1 As shown, Figure 1 A flowchart of a river network water conservancy project scheduling method provided in Embodiment 1 of this application is shown, wherein the method includes steps S101 to S105:

[0097] S101: Obtain basic river network data, determine key control points and their safety thresholds based on the basic river network data, and establish a joint probability distribution model of uncertain input parameters.

[0098] Specifically, multi-source basic data for river network hydrodynamic simulation and water project scheduling is acquired. This multi-source basic data includes at least rainfall data, water level data, flow data, water project operating parameters, river cross-section data, and river network topology data. A unified preprocessing operation is performed on this multi-source basic data, encompassing missing data repair, anomaly identification and removal, time scale unification, spatial matching mapping, and unit unification to ensure consistency in spatiotemporal dimensions and measurement standards. After data preprocessing, a set of key control points for the river network system is constructed based on the river network hydrodynamic characteristics and engineering scheduling requirements. Key control points are river cross-sections or engineering nodes that significantly impact scheduling objectives such as flood control and water supply. A corresponding safety threshold is set for each key control point, serving as the core criterion for subsequent risk assessment and scheduling constraint construction.

[0099] Based on the operational characteristics of the river network system, a set of uncertain input parameters affecting the hydrodynamic response and water engineering scheduling results of the river network is identified, including parameters such as upstream inflow, channel roughness, downstream boundary water level, gate flow coefficient, and scheduling execution deviation. Adaptive probability distribution models are established for each uncertain input parameter, with distribution types including normal, log-normal, empirical, kernel density, or mixed distributions. The distribution parameters are obtained from historical observation data and real-time monitoring data through maximum likelihood estimation. Furthermore, the temporal correlation structure, spatial correlation structure, or multi-source coupled correlation structure among the uncertain input parameters is constructed, and parameterized representation is completed using the covariance matrix. Combined with physical constraints, a joint stochastic input model is formed. The joint random input model is stored in a versioned parameter file format, which contains core information such as the marginal distribution parameters, covariance matrix, and physical constraint descriptions of each parameter. During scene generation, relevant random variables are generated based on the covariance matrix, and the joint random input scene is obtained through inverse marginal distribution transformation. At the same time, physical constraint verification and correction are performed on the generated results to ensure that the scene conforms to actual engineering laws. When the joint random input model iterates due to data updates, a new version identifier and corresponding parameter file are automatically generated. Subsequent scene generation processes read the file according to the latest version identifier, realizing synchronous effect of model updates and full-process traceability.

[0100] S102: Generate a random scenario set based on the joint probability distribution model, and obtain the response results of the key control points in each scenario through hydrodynamic simulation.

[0101] Specifically, based on the safety requirements of flood control and water supply in the engineering scheduling, differentiated target safety satisfaction levels are set for each key control point, serving as a benchmark for probabilistic constraints and risk assessment. Sobol low-discrepancy sequences are used as quasi-Monte Carlo sampling sequences, and an initial random scenario set is generated based on the joint probability distribution model to improve sampling uniformity and efficiency. When the risk assessment module outputs an additional sampling request, the existing scenario set is expanded according to the same sampling rules to supplement new random scenarios, and the updated scenario set is synchronously provided to the hydrodynamic simulation and risk assessment stages.

[0102] For each random scenario in the set of random scenarios, the uncertain parameters within the scenario are mapped to the input conditions of the river network hydrodynamic model, including upstream inflow boundaries, downstream water level boundaries, channel roughness, and engineering operation parameters. The river network hydrodynamic model is run to perform time-series calculations on each scenario, obtaining the hydrodynamic response results such as water level and flow rate at each key control point during the simulation period. Event-level risk indicators related to safety thresholds are extracted from the response results, specifically including maximum water level indicators, duration of exceedance indicators, and peak arrival time indicators, providing basic data for subsequent risk assessment. To reduce the computational cost of multi-scenario hydrodynamic simulation, state-based hot start and parallel batch computing are employed during the simulation process, improving overall computational efficiency without affecting the accuracy of risk assessment.

[0103] S103: Based on the response results, count the sample frequency of the key control points that satisfy the security threshold, and calculate the lower confidence bound of the sample frequency at a given confidence level. The lower confidence bound being no less than the preset target security level is used as the condition for satisfying the probabilistic security constraint.

[0104] Specifically, based on the hydrodynamic response results of key control points under various random scenarios, empirical distribution functions of indicators such as water level and flow rate at key control points at various times are constructed to quantify the distribution patterns of these indicators within the sample range. An event in which a key control point meets a safety threshold at a certain time is defined as a binomial random event. The frequency of this event occurring in all scenarios is counted, and the sample frequency of key control points meeting the safety threshold, i.e., the sample safety satisfaction rate, is calculated.

[0105] Using a default confidence level (e.g., 0.95), the Clopper-Pearson exact confidence interval method with binomial distribution proportions is employed to calculate the statistical confidence interval based on the sample safety satisfaction rate. This yields the lower and upper confidence bounds of the satisfaction rate. This method ensures reliable coverage of the confidence interval under small sample conditions, reducing the probability of false positives. The lower confidence bound of the safety satisfaction rate of key control points is not lower than the preset target safety level, serving as the core condition for satisfying the probabilistic safety constraint of that control point. This enables the determination of the statistical significance of probabilistic safety constraints under finite sample conditions, avoiding safety misjudgments caused by estimation solely based on sample frequency points.

[0106] S104: When the confidence lower bound is lower than the preset target safety level or the confidence interval width of the sample frequency exceeds the preset threshold, additional random scenarios are generated and the hydrodynamic simulation and statistical verification steps are repeated until the probabilistic safety constraint is met or the preset sample upper limit is reached.

[0107] Specifically, based on the hydrodynamic response results under various random scenarios, an empirical distribution of risk indicators is constructed. The Clopper–Pearson exact confidence interval method is used to verify the statistical confidence intervals of the risk indicators, determining whether the lower bound of the sample safety satisfaction rate of key control points reaches the preset target safety level. Simultaneously, using the confidence interval width of the sample safety satisfaction rate of key control points as the criterion for sequential sampling, the difference between the upper and lower confidence bounds is calculated as the interval width value, which is then compared with the preset risk assessment accuracy threshold.

[0108] If either the lower confidence bound falls below the preset target safety level or the confidence interval width exceeds the preset accuracy threshold, the system automatically generates an additional sampling request, specifying the target control point set and the required number of additional samples. Following the sampling rules in S102, a new random scenario is generated, and the hydrodynamic simulation steps are re-executed to obtain the response results of the key control points under the new scenario. The process of constructing the empirical distribution of risk indicators, verifying the confidence interval, and validating the additional sampling criteria is repeated. This process is repeated cyclically until the key control points meet the probabilistic safety constraints or the number of scenario samples reaches the preset maximum sample limit, at which point additional sampling stops.

[0109] S105: Under the premise of satisfying the probabilistic safety constraints, solve the optimization model with the water project operation parameters as decision variables, generate a scheduling scheme, and execute the scheduling scheme.

[0110] Specifically, under the premise that all key control points meet probabilistic safety constraints, a water project scheduling optimization model is constructed with water project operation parameters as decision variables. These decision variables include gate opening sequences, pump station start-up and shutdown schemes, and reservoir inflow and outflow control strategies. The scheduling objective is to minimize flood control risk, and can incorporate objectives such as water supply, power generation, and ecological water demand based on actual project needs, constructing a multi-objective scheduling model. The probabilistic safety constraints defined in S103 are formally introduced as constraints in the scheduling model. The model is solved using an optimized solution algorithm to obtain a water project scheduling scheme that balances safety constraints and scheduling objectives.

[0111] The scheduling scheme is executed using a rolling time-domain scheduling method, wherein the rolling scheduling period is preferably set to 1 hour, and the prediction time domain is preferably set to 24 hours. The prediction time domain is set based on the available accuracy range of the accessed meteorological data. Since the error of meteorological data accumulates with the increase of prediction time, 24 hours is used as the main effective prediction window. In the execution of rolling scheduling, only the decision variables within the current scheduling period are executed. In each subsequent rolling period, based on the latest hydrological monitoring data and meteorological forecast data, the marginal distribution parameters of the uncertain input parameters and the relevant structure represented by the covariance matrix are updated online. The scheduling input scenario is regenerated and the scheduling scheme is corrected according to the updated joint random input model. The process from S102 to this step is repeated to realize the online update, closed-loop verification and continuous smooth execution of water project scheduling.

[0112] In an optional implementation, see Figure 2 As shown, Figure 2 The flowchart of a safety threshold determination method provided in Embodiment 1 of this application is shown, wherein the step of determining key control points and their safety thresholds based on the river network basic data includes steps S201~S203:

[0113] S201: Preprocess the basic river network data to obtain preprocessed data.

[0114] Specifically, the preprocessing operation includes five core operations: missing data repair, abnormal data identification and removal, time scale unification, spatial matching mapping, and unit unification. Through standardization, the differences in spatiotemporal dimensions and measurement standards of multi-source river network basic data are eliminated, ensuring data consistency and validity, and providing a standardized data foundation for subsequent selection of key control points and model calculation.

[0115] S202: Based on the preprocessed data, select key control points from river cross-sections, hydraulic structures, hydrological stations, and important protection targets.

[0116] Specifically, river cross-sections should prioritize those that play a core role in the hydrodynamic safety of the river network, such as urban defense control sections, polder drainage control sections, and estuary control sections; hydraulic structures should select core operational nodes of water projects such as gates and pumping stations; hydrological stations should select standardized monitoring sites with long-term monitoring and complete and effective data; and important protection targets should select hydrological points around areas significantly affected by hydrological conditions, such as urban built-up areas, farmland protection areas, and industrial parks.

[0117] Specifically, based on the hydrological characteristics and engineering scheduling value of the above four types of points, a set of key control points for forming the river network system is constructed, denoted as... Each key control point within the set is distinguished by a unique identifier, corresponding to its specific spatial location and engineering attributes within the river network.

[0118] S203: Set a corresponding safety threshold for each of the aforementioned critical control points.

[0119] Specifically, combining the core objectives of engineering scheduling such as flood control, water supply, power generation, and ecological water demand, safety thresholds for key hydrological indicators such as water level, flow rate, and reservoir capacity are set for each critical control point. The water level safety threshold is denoted as... , (This refers to the key control point number).

[0120] The safety threshold is set with reference to historical hydrological extremes, engineering design standards, regional protection requirements, etc., and can be dynamically adjusted according to real-time operating conditions. This threshold serves as the core criterion for judging whether hydrological indicators exceed limits in subsequent risk assessments, and is also the basis for constructing probabilistic safety constraints.

[0121] In an optional implementation, see Figure 3 As shown, Figure 3 The flowchart of a response result generation method provided in Embodiment 1 of this application is shown, wherein the step of generating a random scene set based on the joint probability distribution model and obtaining the response results of the key control point under each scene through hydrodynamic simulation includes steps S301 to S304:

[0122] S301: The joint probability distribution model is sampled using a quasi-Monte Carlo low-discrepancy sequence to obtain multiple random scenarios, forming an initial random scenario set.

[0123] Specifically, Sobol low-difference sequences were selected as quasi-Monte Carlo sampling sequences, and target safety satisfaction levels were preset for each critical control point based on engineering scheduling safety requirements. ( As a critical control point At any moment (Permissible over-limit risk probability), based on the joint probability distribution model of uncertain input parameters. Perform efficient sampling.

[0124] First, identify the set of uncertain input parameters that affect the hydrodynamic response and scheduling results of the river network, and construct an uncertainty parameter vector: ,in For upstream inflow, For the river channel roughness, This is the downstream boundary water level. The gate flow coefficient, The scheduling execution deviation parameter.

[0125] For joint probability distribution Sampling is performed to generate the first A random scenario: ,in For the first An uncertain scenario, The total number of samples in the current scene; the initial random scene set is composed of all random scenes, denoted as . .

[0126] Before sampling, a probability distribution model and related structure need to be constructed for the uncertainty parameters. Establish a distribution model based on parameters of uncertainty: ,in It is a marginal distribution type (normal / log-normal / empirical distribution / kernel density / mixed distribution). Let be the distribution parameter vector. The total number of categories of uncertainty parameters; simultaneously constructing temporal, spatial, and multi-source coupled correlation structures, with temporal continuity using a first-order autoregressive model: ,in The time correlation coefficient, This is a random disturbance term.

[0127] S302: For each random scene in the initial random scene set, the sampled random scene parameters are mapped to the boundary conditions and model parameters of the hydrodynamic model to obtain the mapped input conditions.

[0128] Specifically, let the first The uncertainty parameter vector corresponding to each random scenario is: By using a dedicated parameter mapping function, each parameter is mapped to the boundary conditions and model parameters of the hydrodynamic model, thereby achieving a precise conversion of uncertain parameters into model inputs.

[0129] Mapping of upstream and tributary inflow boundary conditions: ,in For the first In each scenario, at any moment The inflow process, For the inflow parameter mapping function, For the first Inflow uncertainty parameters under various scenarios; downstream water level boundary condition mapping: ,in For the first In each scenario, at any moment The downstream water level process, This is a mapping function for downstream water level parameters. For the first Downstream boundary uncertainty parameters in a given scenario.

[0130] Channel roughness parameter mapping: ,in For the first River in a certain scenario The roughness parameter, For roughness parameter mapping function, For the first The roughness uncertainty parameter of the river channel r in a given scenario. Represents the river channel; Engineering operation parameter mapping: ,in For the first In each scenario, at any moment Engineering operation control parameters, This is a mapping function for engineering operation parameters. For the first Scheduling execution deviation parameters in various scenarios The corresponding sample value.

[0131] S303: Run the river network hydrodynamic model based on the mapped input conditions to obtain the water level sequence and flow sequence of each key control point during the simulation period.

[0132] Specifically, for each random scenario and its mapped input conditions, the time series calculation of the river network hydrodynamic model is performed to solve the dynamic changes in river level and flow, and output the time series data of water level and flow of each key control point throughout the simulation period.

[0133] The water level sequence of the key control points is denoted as: ,in As a critical control point In the In a random scenario, at a certain time The water level and flow rate sequence are denoted as: ,in As a critical control point In the In a random scenario, at a certain time Traffic.

[0134] To reduce the overall computational cost of multi-scenario hydrodynamic simulation, engineering methods such as state-based hot start and parallel batch computing are adopted in the model calculation process to improve the overall computational efficiency without affecting the accuracy of risk assessment.

[0135] S304: Extract the response results of each key control point from the water level sequence and flow rate sequence.

[0136] Specifically, taking the water level sequence of key control points as the core, event-level risk indicators are extracted for risk assessment and scheduling constraint construction. At the same time, combined with the characteristic parameters of the flow sequence, the hydrodynamic response results of key control points are formed. These results serve as the basic input for subsequent risk assessment and probabilistic constraint verification.

[0137] Extract the maximum water level index: ,in As a critical control point In the The maximum water level in a random scenario To simulate the total duration; extract the over-limit duration index: ,in As a critical control point In the The duration of water level exceeding the limit in a random scenario. This is an indicator function; it takes the value 1 if the condition is met, and 0 if the condition is not met.

[0138] Extracting peak arrival time metrics: ,in As a critical control point In the The peak water level arrival time under random scenarios; the above three types of risk indicators characterize the hydrodynamic risk status of key control points from different dimensions, providing core data for subsequent statistical analysis.

[0139] In an optional implementation, see Figure 4 As shown, Figure 4 The flowchart of a sample frequency generation method provided in Embodiment 1 of this application is shown, wherein the step of statistically analyzing the sample frequency of the key control point satisfying the security threshold based on the response result includes steps S401 to S403:

[0140] S401: For each key control point, determine whether its response result in each random scenario meets the security threshold, and obtain the satisfaction status in each scenario.

[0141] Specifically, key control points At any moment Events that meet the safety threshold are considered binomial random events, and an indicator variable for this event is defined as follows: ,in As a critical control point At any moment , No. A state indicator variable for a random scenario, where 1 indicates that the safety threshold is met and 0 indicates that the limit is exceeded.

[0142] Based on the hydrodynamic response results under various random scenarios, and in accordance with the above indicator variable formula, the satisfaction status judgment of all key control points under all times and all random scenarios is completed one by one, forming a complete status judgment matrix to ensure no omissions and no misjudgments.

[0143] S402: Count the number of scenarios where the satisfied state is "yes" to obtain the number of satisfied scenarios.

[0144] Specifically, for the same key control point At the same time Satisfaction state indicator variables for all random scenarios To perform summation, i.e. The total number of times the critical control point meets the safety threshold at that moment is calculated, and the summation process covers all of the initial random scenario set. One sample.

[0145] S403: The ratio of the number of times the condition is satisfied to the total number of scenarios is used as the sample frequency.

[0146] Specifically, the number of times the condition is satisfied is divided by the total number of samples in the initial random scene set. The calculated ratio is the sample safety satisfaction rate of the critical control point at that moment, which is also known as the sample frequency. The calculation formula is as follows: ,in As a critical control point At any moment The sample safety satisfaction rate is the core foundational data for subsequent calculations of statistical confidence intervals and determination of probability safety constraints.

[0147] In an optional implementation, see Figure 5 As shown, Figure 5 The flowchart of a confidence lower bound determination method provided in Embodiment 1 of this application is shown, wherein the step of calculating the confidence lower bound at a given confidence level based on the sample frequency includes steps S501 to S503:

[0148] S501: Obtain the sample frequency and the corresponding total number of random scenarios.

[0149] Specifically, extract the key control points calculated in step S403. At any moment Sample safety satisfaction rate At the same time, obtain the total number of samples in the current random scenario set used for risk assessment. This ensures that the two sets of data match the corresponding key control points and simulation times one by one, without any overlap or confusion.

[0150] Based on the water level response results under various random scenarios, key control points are first constructed. At any moment The empirical distribution function of water level provides a distribution basis for the calculation of confidence intervals, and the formula is: ,in Let be the empirical distribution function of water level. For water level variables, This is an indicator function.

[0151] S502: Using the binomial distribution ratio precise confidence interval construction method, the confidence interval is calculated based on the sample frequency, the total number of random scenarios and the given confidence level.

[0152] Specifically, the Clopper–Pearson exact confidence interval construction method is selected. This method can guarantee the reliable coverage of confidence intervals under small sample conditions, effectively reducing the probability of false positives; and the sample safety satisfaction rate is... Total number of random scenes As input parameters, combined with the given confidence level (The default value is 0.95) is used for calculation, where β is the confidence interval significance parameter.

[0153] The calculated statistical confidence interval corresponding to the sample safety satisfaction rate is denoted as: ,in The lower bound of the safety satisfaction rate, This serves as the upper bound for the safety satisfaction rate. Before calculating the confidence interval, auxiliary risk indicators such as the probability of exceeding the limit, quantile water level, and conditional risk value can be calculated based on the empirical distribution of water levels to improve the risk assessment system.

[0154] Estimation of out-of-limit probability samples: ,in As a critical control point At any moment Probability of water level exceeding limit; quantile water level: For a given target risk level The corresponding sample quantile level.

[0155] Conditional risk value: ,in As a critical control point At any moment Target risk level The risk value of the water level condition is E[·], which is the mathematical expectation.

[0156] S503: Use the lower limit of the confidence interval as the lower confidence bound.

[0157] Specifically, extract the lower limit of the statistical confidence interval calculated in step S502. Use it as a key control point At any moment The lower bound of the sample safety satisfaction rate is the core indicator for determining whether the probabilistic safety constraints are met, and it is also one of the criteria for judging the sequential sampling mechanism.

[0158] Simultaneously, the confidence interval width is calculated as another criterion for sequential sampling, and the formula is: ,in As a critical control point The width of the confidence interval for the safety satisfaction rate, if ( If the preset risk assessment accuracy threshold is not met, the current sample size is determined to be insufficient, triggering a sample increase request.

[0159] In an optional implementation, see Figure 6 As shown, Figure 6 The flowchart of a scheduling scheme generation method provided in Embodiment 1 of this application is shown. The step of solving an optimization model with water project operation parameters as decision variables to generate a scheduling scheme, under the premise of satisfying the probabilistic security constraints, includes steps S601-S604:

[0160] S601: Construct an objective function by using at least one of the following as decision variables: gate opening sequence, pump station start-up and shutdown scheme, and reservoir inflow and outflow.

[0161] Specifically, the scheduling time domain is defined as follows: ,in Given the total number of time nodes within the scheduling time domain; construct a vector of scheduling decision variables: ,in For a moment The gate opening, For a moment The pump station's start / stop status. For a moment The inflow and outflow of the reservoir group.

[0162] With minimizing flood control risk as the core scheduling objective, a flood control risk objective function is constructed by introducing risk weight coefficients for key control points, as shown in the formula: ,in For the objective function of flood control risk, As a critical control point Risk weighting coefficients (assigned based on the importance of control points). To take the larger function, Let be the mathematical expectation.

[0163] Based on the actual needs of the project, objectives such as water supply security, power generation benefits, and ecological water demand can be integrated into the flood control risk objective to construct a multi-objective scheduling function. At the same time, the project operation cost and scheduling stability are comprehensively considered to make the objective function more in line with the actual project scheduling needs.

[0164] S602: Use the probabilistic safety constraints as constraints for the optimization model.

[0165] Specifically, the water level safety requirements at key control points are first transformed into the basic form of probabilistic safety constraints: ,in For probability operators, As a critical control point At any moment The target safety level is met.

[0166] To reduce the risk of misjudgment under limited samples, the probabilistic safety constraint is conservatively replaced by a one-sided confidence lower bound of the sample safety satisfaction rate, forming the actual constraint conditions of the optimization model: ,in In the scheduling scheme Below, key control points At any moment The safety satisfaction rate confidence lower bound can be included; at the same time, it can be incorporated into engineering hard constraints such as gate opening limit, pump station operating power, and reservoir capacity range to improve the constraint system.

[0167] S603: Based on the objective function and the constraints, solve the optimization model using mathematical programming or intelligent optimization algorithms to obtain the optimized decision variable values.

[0168] Specifically, mathematical programming algorithms can use classic optimization algorithms such as linear programming, nonlinear programming, and dynamic programming, which are suitable for scenarios where the objective function and constraints are explicit expressions; intelligent optimization algorithms can use heuristic algorithms, which are suitable for complex nonlinear scheduling scenarios involving river network hydrodynamic coupling.

[0169] With minimizing the objective function as the optimization direction, and under the premise of satisfying all probabilistic safety constraints and engineering hard constraints, the scheduling optimization model is solved to obtain the optimal values ​​of the scheduling decision variables at each time node, denoted as: ,in As a critical control point The corresponding optimal scheduling decision variable vector, For the comprehensive scheduling objective function, This is the set of scheduling decision variables.

[0170] S604: Generate an executable water engineering scheduling scheme based on the optimized decision variable values.

[0171] Specifically, the optimal scheduling decision variables obtained from the optimization solution, such as the gate opening sequence, pump station start / stop status, and reservoir inflow / outflow, will be optimized. The timing values ​​are combined with the operational requirements of the engineering site to transform them into execution instructions that can be directly applied to engineering operations.

[0172] The generated scheduling scheme is based on the optimal scheduling decision variables. The plan includes core information such as time nodes, operation objects, operation parameters, and execution requirements, such as the opening value of a gate at a certain moment, the start and stop time of a pumping station, and the inflow and outflow of a reservoir, to ensure the feasibility of the scheduling plan. At the same time, the plan will also include the risk assessment results of each key control point to provide quantitative reference for scheduling decisions.

[0173] In an optional implementation, the method further includes at least one of the following:

[0174] The scheduling scheme is executed according to the rolling time domain to obtain real-time monitoring data during the execution process.

[0175] Specifically, a rolling time-domain scheduling method is used to execute based on The generated scheduling scheme preferably sets the rolling scheduling cycle to 1 hour and the prediction time domain to 24 hours. The prediction time domain is set based on the available accuracy range of the accessed meteorological data. Since the error of meteorological data accumulates as the prediction time increases, 24 hours is used as the main effective prediction window.

[0176] During the rolling scheduling execution process, only the decision variables within the current scheduling period are executed. 1. At the same time, through terminals such as hydrological monitoring stations and engineering operation monitoring equipment, real-time monitoring data on rainfall, water level, flow rate, gate opening, and pump station operation status are continuously collected to ensure the timeliness, completeness and accuracy of the data.

[0177] The joint probability distribution model is updated online based on the real-time monitoring data to obtain the updated joint probability distribution model.

[0178] Specifically, based on the collected real-time monitoring data, the marginal distribution parameters and related structures of the covariance matrix of the uncertain input parameters are updated online. The update methods include sliding time window statistical update, recursive maximum likelihood estimation, Bayesian update or particle filter update, and the appropriate method can be selected according to the data characteristics and engineering requirements.

[0179] When the uncertainty parameter is modeled using a normal distribution, the distribution parameter is updated according to a recursive formula, with the initial distribution being... ,in This is the initial statistical mean. Initial discreteness; obtaining new observation samples Then, the updated mean: Updated variance: .

[0180] In the above normal distribution update formula, The mean of the new observed sample. The variance of the observation error. The number of new observation samples; the updated joint random input model is stored in a versioned parameter file, which includes marginal distribution parameters, covariance matrix, and physical constraint description, enabling traceability of model updates.

[0181] Based on the updated joint probability distribution model, the step of generating the random scene set is returned, forming a closed-loop operation.

[0182] Specifically, based on the updated joint probability distribution model, quasi-Monte Carlo sampling is performed again using Sobol low-discrepancy sequences to generate a new set of random scenes, which replaces the original set of scenes in subsequent calculations.

[0183] Based on a new set of random scenarios, the process of hydrodynamic simulation, risk assessment and probability constraint verification, and scheduling optimization solution is repeated to optimize the scheduling decision variables and generate new scheduling schemes. This forms a closed-loop operation mechanism of "prediction-evaluation-scheduling-update-reprediction", which enables the scheduling scheme to be dynamically corrected in response to changes in inflow, boundary conditions, and engineering operation status.

[0184] Record the risk level, sample size, statistical confidence interval, and execution effect of each scheduling decision to form a traceable scheduling decision log.

[0185] Specifically, the core information of each scheduling decision is recorded uniformly to form a standardized scheduling decision log, which is recorded as follows: ,in For optimal decision variables, To ensure the safety satisfaction rate of the sample, To have faith in the lower realm, The number of random scenes. Risk level.

[0186] The scheduling decision log also needs to record scheduling execution time, real-time monitoring data, actual project execution effect, adverse scenario information, etc., to provide traceable and complete data support for subsequent model correction, scheduling strategy optimization, and project operation evaluation.

[0187] When real-time monitoring data is missing or delayed, online inference is performed based on the established joint probability distribution model to obtain inferred data, and the continuity of scheduling operation is maintained based on the inferred data.

[0188] Specifically, when abnormal situations such as missing real-time monitoring data or transmission delays occur, the missing / delayed monitoring data is inferred and supplemented online based on the probability distribution characteristics and parameter correlation structure of the established joint probability distribution model, so as to obtain inferred data that conforms to the data distribution law.

[0189] Using inferred data as model input maintains the continuity of hydrodynamic simulation and scheduling calculations, while employing a confidence lower bound check function for fault tolerance; if data quality fluctuations cause statistical accuracy to fall below the threshold... Or believe in the lower realm Below the target safety satisfaction level The system instantly triggers event responses, generates random scenario sets, performs scenario sampling, and re-solves the scheduling decision variables. 1. And recalculate subsequent processes to ensure the safety and stability of scheduling operations.

[0190] Example 2

[0191] See Figure 7 As shown, Figure 7 This paper shows a schematic diagram of a river network water conservancy project scheduling device according to Embodiment 2 of this application, wherein the device includes:

[0192] The model building module 701 is used to acquire basic river network data, determine key control points and their safety thresholds based on the basic river network data, and establish a joint probability distribution model of uncertain input parameters.

[0193] The response result generation module 702 is used to generate a random scenario set based on the joint probability distribution model, and obtain the response results of the key control point under each scenario through hydrodynamic simulation.

[0194] The condition determination module 703 is used to statistically analyze the sample frequency of the key control points satisfying the security threshold based on the response results, and calculate the lower confidence bound of the sample frequency at a given confidence level, with the lower confidence bound not being lower than the preset target security level as the condition for satisfying the probabilistic security constraint.

[0195] The statistical verification module 704 is used to generate additional random scenarios and repeat the hydrodynamic simulation and statistical verification steps when the confidence lower bound is lower than the preset target safety level or the confidence interval width of the sample frequency exceeds the preset threshold, until the probability safety constraint is met or the preset sample upper limit is reached.

[0196] The scheduling scheme execution module 705 is used to solve an optimization model with water project operation parameters as decision variables, generate a scheduling scheme, and execute the scheduling scheme, under the premise of satisfying the probabilistic safety constraints.

[0197] In an optional implementation, determining the key control points and their safety thresholds based on the river network infrastructure data includes:

[0198] The river network basic data is preprocessed to obtain preprocessed data;

[0199] Based on the preprocessed data, key control points are selected from river cross sections, hydraulic structures, hydrological stations, and important protection targets;

[0200] Set a corresponding safety threshold for each of the aforementioned critical control points.

[0201] In an optional implementation, the step of generating a random scenario set based on the joint probability distribution model and obtaining the response results of the key control points under each scenario through hydrodynamic simulation includes:

[0202] The joint probability distribution model is sampled using quasi-Monte Carlo low-discrepancy sequences to obtain multiple random scenarios, forming an initial random scenario set;

[0203] For each random scene in the initial random scene set, the sampled random scene parameters are mapped to the boundary conditions and model parameters of the hydrodynamic model to obtain the mapped input conditions.

[0204] Based on the mapped input conditions, the river network hydrodynamic model is run to obtain the water level sequence and flow sequence of each key control point during the simulation period.

[0205] The response results of each key control point are extracted from the water level sequence and flow rate sequence.

[0206] In an optional implementation, the step of statistically analyzing the sample frequency of the critical control points satisfying the security threshold based on the response results includes:

[0207] For each key control point, determine whether its response result in each random scenario meets the security threshold, and obtain the satisfaction status in each scenario.

[0208] Count the number of scenarios where the satisfaction state is "yes" to obtain the satisfaction count;

[0209] The ratio of the number of times the condition is satisfied to the total number of scenarios is used as the sample frequency.

[0210] In an optional implementation, calculating the lower confidence bound of the sample based on the sample frequency at a given confidence level includes:

[0211] Obtain the sample frequency and the corresponding total number of random scenarios;

[0212] A method for constructing precise confidence intervals using binomial distribution proportions is employed, based on the sample frequency, the total number of random scenarios, and the given confidence level, to calculate the confidence intervals.

[0213] The lower limit of the confidence interval is used as the lower confidence bound.

[0214] In an optional implementation, the step of solving an optimization model with water project operating parameters as decision variables to generate a scheduling scheme, under the premise of satisfying the probabilistic safety constraints, includes:

[0215] An objective function is constructed by using at least one of the following as decision variables: gate opening sequence, pump station start-up and shutdown scheme, and reservoir inflow and outflow.

[0216] The probabilistic safety constraints are used as constraints on the optimization model.

[0217] Based on the objective function and the constraints, the optimization model is solved using mathematical programming or intelligent optimization algorithms to obtain the optimized values ​​of the decision variables.

[0218] Based on the optimized decision variable values, an executable water project scheduling scheme is generated.

[0219] In an optional implementation, the apparatus further includes an execution monitoring module for performing at least one of the following methods:

[0220] The scheduling scheme is executed in a rolling time domain to obtain real-time monitoring data during the execution process;

[0221] The joint probability distribution model is updated online based on the real-time monitoring data to obtain the updated joint probability distribution model.

[0222] Based on the updated joint probability distribution model, the step of generating the random scene set is returned, forming a closed-loop operation;

[0223] Record the risk level, sample size, statistical confidence interval, and execution effect of each scheduling decision to form a traceable scheduling decision log;

[0224] When real-time monitoring data is missing or delayed, online inference is performed based on the established joint probability distribution model to obtain inferred data, and the continuity of scheduling operation is maintained based on the inferred data.

[0225] Example 3

[0226] Based on the same application concept, see [link / reference] Figure 8 As shown, Figure 8 This illustration shows a structural schematic diagram of a computer device provided in Embodiment 3 of this application, wherein, as shown... Figure 8 As shown, the computer device 800 provided in Embodiment 3 of this application includes:

[0227] The computer device 800 includes a processor 801, a memory 802, and a bus 803. The memory 802 stores machine-readable instructions that can be executed by the processor 801. When the computer device 800 is running, the processor 801 communicates with the memory 802 through the bus 803. When the machine-readable instructions are executed by the processor 801, the steps of the river network water engineering scheduling method shown in Embodiment 1 are performed.

[0228] Example 4

[0229] Based on the same concept, this application also provides a computer-readable storage medium storing a computer program, which, when run by a processor, executes the steps of the river network water engineering scheduling method described in any of the above embodiments.

[0230] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system and apparatus described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0231] The computer program product for scheduling river network water conservancy projects provided in this application includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0232] The river network water conservancy project scheduling device provided in this application embodiment can be specific hardware on the equipment or software or firmware installed on the equipment. The implementation principle and technical effects of the device provided in this application embodiment are the same as those in the foregoing method embodiments. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the foregoing method embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.

[0233] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0234] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0235] In addition, the functional units in the embodiments provided in 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.

[0236] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion 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 a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0237] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0238] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A method for scheduling river network water conservancy projects, characterized in that, The method includes: Acquire basic river network data, determine key control points and their safety thresholds based on the basic river network data, and establish a joint probability distribution model of uncertain input parameters. The key control points include river cross sections, hydraulic structures, hydrological stations, and important protection objects. Based on the joint probability distribution model, a random scenario set is generated, and the response results of the key control points under each scenario are obtained through hydrodynamic simulation. Based on the response results, the frequency of samples that the key control points meet the security threshold is statistically analyzed, and the lower bound of the confidence level is calculated based on the sample frequency. The lower bound of the confidence level is not lower than the preset target security level as the condition for satisfying the probabilistic security constraint. When the lower confidence bound is lower than the preset target safety level or the confidence interval width of the sample frequency exceeds the preset threshold, additional random scenarios are generated and the hydrodynamic simulation and statistical verification steps are repeated until the probabilistic safety constraint is met or the preset sample upper limit is reached. Under the premise of satisfying the probabilistic safety constraints, solve the optimization model with water project operation parameters as decision variables, generate a scheduling scheme, and execute the scheduling scheme; The process of generating a random scenario set based on the joint probability distribution model and obtaining the response results of the key control points under each scenario through hydrodynamic simulation includes: For each random scenario in the set of random scenarios, the uncertainty parameters within the random scenario are mapped to the input conditions of the river network hydrodynamic model. The uncertainty parameters include the upstream inflow boundary, the downstream water level boundary, the channel roughness, and the engineering operation parameters. The river network hydrodynamic model is run to perform time-series calculations on the random scenario, and the hydrodynamic response results of each key control point during the simulation period are obtained. The hydrodynamic response results include water level and flow rate.

2. The method according to claim 1, characterized in that, The determination of key control points and their safety thresholds based on the river network basic data includes: The river network basic data is preprocessed to obtain preprocessed data; Based on the preprocessed data, key control points are selected from river cross sections, hydraulic structures, hydrological stations, and important protection targets; Set a corresponding safety threshold for each of the aforementioned critical control points.

3. The method according to claim 1, characterized in that, The process of generating a random scenario set based on the joint probability distribution model and obtaining the response results of the key control points under each scenario through hydrodynamic simulation includes: The joint probability distribution model is sampled using quasi-Monte Carlo low-discrepancy sequences to obtain multiple random scenarios, forming an initial random scenario set; For each random scene in the initial random scene set, the sampled random scene parameters are mapped to the boundary conditions and model parameters of the hydrodynamic model to obtain the mapped input conditions. Based on the mapped input conditions, the river network hydrodynamic model is run to obtain the water level sequence and flow sequence of each key control point during the simulation period. The response results of each key control point are extracted from the water level sequence and flow rate sequence.

4. The method according to claim 1, characterized in that, The step of statistically analyzing the sample frequency of the key control points satisfying the security threshold based on the response results includes: For each key control point, determine whether its response result in each random scenario meets the security threshold, and obtain the satisfaction status in each scenario. Count the number of scenarios where the satisfaction state is "yes" to obtain the satisfaction count; The ratio of the number of times the condition is satisfied to the total number of scenarios is used as the sample frequency.

5. The method according to claim 1, characterized in that, The calculation of the lower confidence bound at a given confidence level based on the sample frequency includes: Obtain the sample frequency and the corresponding total number of random scenarios; A method for constructing precise confidence intervals using binomial distribution proportions is employed, based on the sample frequency, the total number of random scenarios, and the given confidence level, to calculate the confidence intervals. The lower limit of the confidence interval is used as the lower confidence bound.

6. The method according to claim 1, characterized in that, The step of solving an optimization model with water project operation parameters as decision variables, under the premise of satisfying the probabilistic safety constraints, to generate a scheduling scheme includes: An objective function is constructed by using at least one of the following as decision variables: gate opening sequence, pump station start-up and shutdown scheme, and reservoir inflow and outflow. The probabilistic safety constraints are used as constraints on the optimization model. Based on the objective function and the constraints, the optimization model is solved using mathematical programming or intelligent optimization algorithms to obtain the optimized values ​​of the decision variables. Based on the optimized decision variable values, an executable water project scheduling scheme is generated.

7. The method according to claim 1, characterized in that, The method further includes: The scheduling scheme is executed in a rolling time domain to obtain real-time monitoring data during the execution process; The joint probability distribution model is updated online based on the real-time monitoring data to obtain the updated joint probability distribution model. Based on the updated joint probability distribution model, the step of generating the random scene set is returned, forming a closed-loop operation; Record the risk level, sample size, statistical confidence interval, and execution effect of each scheduling decision to form a traceable scheduling decision log; When real-time monitoring data is missing or delayed, online inference is performed based on the established joint probability distribution model to obtain inferred data, and the continuity of scheduling operation is maintained based on the inferred data.

8. A river network water engineering scheduling device, characterized in that, The device includes: The model building module is used to acquire basic river network data, determine key control points and their safety thresholds based on the basic river network data, and establish a joint probability distribution model of uncertain input parameters. The key control points include river cross sections, hydraulic structures, hydrological stations, and important protection objects. The response result generation module is used to generate a random scenario set based on the joint probability distribution model, and obtain the response results of the key control points under each scenario through hydrodynamic simulation. The condition determination module is used to statistically analyze the sample frequency of the key control points satisfying the security threshold based on the response results, and calculate the lower confidence bound of the sample frequency at a given confidence level, with the lower confidence bound not being lower than the preset target security level as the condition for satisfying the probabilistic security constraint. The statistical verification module is used to generate additional random scenarios and repeat the hydrodynamic simulation and statistical verification steps when the confidence lower bound is lower than the preset target safety level or the confidence interval width of the sample frequency exceeds the preset threshold, until the probability safety constraint is met or the preset sample upper limit is reached. The scheduling scheme execution module is used to solve the optimization model with water project operation parameters as decision variables, generate a scheduling scheme, and execute the scheduling scheme, under the premise of satisfying the probabilistic safety constraints. The process of generating a random scenario set based on the joint probability distribution model and obtaining the response results of the key control points under each scenario through hydrodynamic simulation includes: For each random scenario in the set of random scenarios, the uncertainty parameters within the random scenario are mapped to the input conditions of the river network hydrodynamic model. The uncertainty parameters include the upstream inflow boundary, the downstream water level boundary, the channel roughness, and the engineering operation parameters. The river network hydrodynamic model is run to perform time-series calculations on the random scenario, and the hydrodynamic response results of each key control point during the simulation period are obtained. The hydrodynamic response results include water level and flow rate.

9. A computer device, characterized in that, include: The system includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the computer device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, they perform the steps of the river network water engineering scheduling method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the river network water engineering scheduling method as described in any one of claims 1 to 7.

Images