A joint intelligent dispatching system and dispatching method for water conservancy projects in river basins

Abstract

This invention discloses a watershed water project joint intelligent scheduling system and scheduling method for smart water conservancy. The system includes a watershed water project joint scheduling knowledge graph distiller, a watershed water conservancy model set, and a water project joint scheduling plan reasoner. The watershed water project joint scheduling knowledge graph distiller uses a watershed data base and expert instructions to bidirectionally distill the water project joint scheduling operation mode. The watershed water conservancy model set includes project scheduling, watershed simulation, and artificial intelligence modules to serve the watershed water project joint scheduling knowledge graph distillation and water project joint scheduling plan reasoning. The water project joint scheduling plan reasoner establishes a bidirectional mapping relationship between water project scheduling measures and treatment results through a driving feedback mechanism. Based on the scheduling boundary conditions of the current treatment event, it conducts optimization of participating water projects and reasoning on scheduling measures matching. Based on big data, knowledge reasoning, and water conservancy professional models, this invention provides recommendations for watershed water project joint scheduling.

Description

Technical Field

[0001] The present invention relates to a watershed water project joint intelligent dispatching system and dispatching method for smart water conservancy, and belongs to water conservancy and hydropower engineering and water ecological protection. Background Art

[0002] The intelligent scheduling of river basin water projects is a key area in water resources management, balancing multiple objectives, including flood control, water supply, and ecological and environmental protection. Due to the complexity of river basins and the interdependence between projects, traditional scheduling methods often struggle to meet the demands of multi-objective optimization.

[0003] Some publicly available research focuses on flood control, building knowledge graphs to organize data and recommend scheduling rules. For example, by extracting hydraulic engineering object data and historical scheduling cases, flood control scheduling recommendations are generated. However, these recommendations are limited to flood control objectives and fail to consider other integrated management objectives, such as water resources or aquatic ecosystems. Other technologies utilize knowledge graphs for flood forecasting, digitally mapping river networks and building predictive models. This approach is effective for disaster warning, but does not address the actual scheduling decisions of water projects and lacks support for multi-objective optimization. Other publicly available technologies focus on visualizing the relationships between water resource objects, using knowledge graphs to display the relationships between entities such as rivers, lakes, and dams. This approach facilitates data management but does not directly participate in scheduling decisions, making it difficult to directly respond to real-time scheduling needs. Some technical approaches target optimized reservoir scheduling for flood control, leveraging historical data and matching techniques to generate scheduling plans. However, these methods are often limited to single reservoirs and do not consider the coordinated scheduling of multiple projects, making them difficult to adapt to the overall needs of complex river basins.

[0004] Based on the above analysis, existing technologies have made some breakthroughs in single objectives (such as flood control) or specific scenarios (such as reservoirs), but lack a comprehensive system that can simultaneously address multiple scheduling objectives, integrate multiple models, and achieve intelligent optimization. There is an urgent need for a comprehensive solution for the joint scheduling of water projects that takes multiple objectives into account. By integrating knowledge graphs and model sets, a framework for multi-objective collaborative optimization is provided. Through a water project joint scheduling plan reasoner, especially a driven feedback mechanism between scheduling measures and treatment results, the optimal selection of participating water projects and the matching reasoning of joint scheduling plans under changing conditions are carried out, filling the gaps in existing technologies in real-time response. Through expert knowledge and data-driven two-way distillation, a deep fusion of expert experience and data analysis is achieved, surpassing the limitations of existing technologies in decision support approaches. Summary of the Invention

[0005] Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides a watershed water project joint intelligent scheduling system and scheduling method for smart water conservancy. Based on multi-source data and expert knowledge, it generates a water project scheduling operation mode that adapts to the characteristics of complex watersheds, and infers water project scheduling plans through a driving feedback mechanism to achieve efficient decision support, thereby deeply integrating data-driven scientific analysis with expert-guided experience and knowledge.

[0006] Technical Solution: To solve the above technical problems, the present invention provides a watershed water project joint intelligent scheduling system for smart water conservancy, including a watershed water project joint scheduling knowledge graph distiller, a watershed water conservancy model set, and a water project joint scheduling plan reasoner;

[0007] The knowledge graph distiller for joint scheduling of water projects in the basin, based on the basin data base and expert instructions, analyzes the impact mechanism of water project scheduling on various scheduling targets, and constructs the operation mode of joint scheduling of water projects in the basin through bidirectional distillation;

[0008] The watershed water conservancy model set includes an engineering scheduling module, a watershed simulation module, a mathematical statistics module, and an artificial intelligence module. It applies water conservancy mathematical models, statistical models, and artificial intelligence models to serve the knowledge graph distillation and water project joint scheduling plan reasoning of the watershed water project. Among them, the engineering scheduling module, watershed simulation module, mathematical statistics module, and artificial intelligence module all use existing modules.

[0009] The water project joint scheduling plan reasoner constructs a two-way mapping relationship of the driving feedback mechanism between water project scheduling measures and treatment results based on the measured extreme event archives of basin floods, the measured event archives of different occurrence frequencies, and the simulated combined operating condition data. When responding to the current treatment event, the two-way mapping relationship of the driving feedback mechanism is reversely applied to carry out the optimization of the participating water projects and the matching reasoning of the participating water project scheduling measures.

[0010] Preferably, the basin data base comprises real-time monitoring data, compiled data, statistical bulletins, water project design reports and scheduling regulations, and characteristic water levels of control sections of the basin, related to flood control, water resources, and water ecological scheduling objectives of the basin water projects, including data archives corresponding to measured extreme flood events in the basin and measured events with different occurrence frequencies. The data archives include scheduling boundary conditions and disposal sections. The scheduling boundary conditions are the initial values ​​of water levels and flows at hydrological stations in the basin before water project scheduling, as well as flood process vectors, peak flows, peak occurrence times, and flood volumes of the basin water projects and downstream intervals.

[0011] The expert instructions of the water project joint scheduling knowledge graph distiller are instructions from water conservancy industry experts for the joint scheduling of water projects, including participating water projects, scheduling measures, and scheduling constraints.

[0012] Preferably, the engineering scheduling module includes a water engineering control application model;

[0013] The water project control application model applies the water balance principle to carry out continuous calculations at a preset time step during the scheduling period to obtain the upstream water level, storage capacity and discharge flow process of the water project; at each calculation step, relying on the water level-volume curve, water level-discharge capacity curve and water level-controlled discharge curve of the water project, the real-time water storage capacity, upper and lower limits of the discharge capacity of the water project are analyzed, and the water level and discharge flow requirements in the water project scheduling regulations are faced. The period discharge flow is calculated in combination with the period inflow flow, and the upstream water level and storage capacity at the end of the water project period are deduced based on the water balance.

[0014] Preferably, the watershed simulation module is used to simulate and analyze the watershed, including a watershed hydrological model and a hydrodynamic model;

[0015] The basin hydrological model solves the hydrological equations for precipitation distribution, evapotranspiration, runoff generation, slope runoff, river confluence, and groundwater movement based on basin meteorological data, topographic characteristics, and soil types. River confluence considers evaporation and seepage losses, water abstraction along the river, and inter-regional confluence, and outputs the basin outlet runoff process, flood flow, and water level process.

[0016] The hydrodynamic model solves shallow water equations for the entire or partial flood storage space of the basin to simulate water movement, and calculates the hydraulic information of any area based on spatial water inflow, including water level, water depth, flow rate, flow velocity, and flow direction.

[0017] A scheduling method for a watershed water project joint intelligent scheduling system for smart water conservancy, comprising the following steps:

[0018] S1. Collect the data base of a certain watershed;

[0019] S2. Obtain instructions from experts on joint dispatch of water projects in the basin;

[0020] S3. Knowledge Graph Distiller for Joint Scheduling of River Basin Water Projects: Based on the river basin data base and the river basin hydraulic model set, this tool quantifies the impact of water project scheduling on flood control, water resources, and water ecological scheduling objectives, and establishes overall requirements for the joint scheduling of water projects included in the joint scheduling scope. It also conducts semantic recognition of expert instructions for joint scheduling of river basin water projects, including participating water projects, scheduling measures, and scheduling constraints, and formulates a joint scheduling operation mode for water projects based on the overall requirements for joint scheduling of water projects.

[0021] S4. Constructing a water project joint dispatch plan reasoner: Based on the flood extreme event archives, the archives of events with different frequencies, and the simulated combined operating condition data in the basin data base, and according to the basin water project joint dispatch operation mode and basin hydraulic model set, a set of water project joint optimized dispatch plans is generated. This establishes a bidirectional mapping relationship between water project dispatch measures and the driving feedback mechanism of the treatment results.

[0022] S5. Use the water project joint scheduling plan reasoner to respond to the current disposal event and carry out the optimization of the participating water projects and the matching reasoning of the scheduling measures: obtain the scheduling boundary conditions of the current disposal event, the user-selected disposal section and the user-set disposal effect, and use the similarity analysis model of the mathematical statistics module to match the similar flood scenarios with the scheduling boundary conditions and disposal sections of the current disposal event in the measured extreme event archives, the measured event archives with different occurrence frequencies and the simulated combined working condition data. Based on the bidirectional mapping relationship between the water project scheduling measures and the disposal effect of the similar scenarios, the bidirectional mapping relationship of the driving feedback mechanism is reversely applied according to the user-set disposal effect to match the pre-discharge drop level amplitude of the water projects included in the joint scheduling scope, among which the water projects with the pre-discharge drop level amplitude greater than zero are regarded as the optimization results of the participating water projects, and the corresponding pre-discharge drop level amplitude is regarded as the scheduling measure matching reasoning result. The water project joint scheduling plan reasoner outputs the optimization results of the participating water projects and the scheduling measure matching reasoning results.

[0023] Preferably, step S3 comprises the following steps:

[0024] S31. Extract long-term inflow data for each water project's catchment area and downstream sections from the basin data base, water resource demand and water withdrawal ratio data for production, domestic, ecological, and agricultural water use outside the river channel, ecological water demand data at different locations within the river channel, and water project replenishment loss coefficients; extract the scheduling procedures, characteristic parameters, and curves for each water project, including allowable water level fluctuations, water level-volume curves, water level-discharge capacity curves, and water level-controlled discharge curves; and extract the warning water level and guaranteed water level for the control section of the downstream storage space of each water project.

[0025] S32. Discretize the maximum possible amplitude of the pre-discharge water level drop of each water project compared to the flood limit water level during the flood season, enumerate the combinations of water level drop amplitudes among multiple water projects, construct a water project joint pre-discharge scheduling operating condition set, and perform simulation calculations for the water project and downstream river channel for each operating condition in the operating condition set;

[0026] S33. Carry out simulation calculations of the water project and downstream river channel for the joint pre-discharge scheduling conditions of each water project: Use the water project control application model to carry out long-term continuous calculations to obtain long-term data on the discharge flow, upstream water level, and storage capacity of each water project. Combined with the basin hydrological model or hydrodynamic model of the basin simulation module, carry out water flow routing of the hydraulic elements of the downstream storage space of the water project. The water flow routing takes into account the water replenishment loss of the water project, the incoming water along the section, the water intake and return of water outside the river channel. The calculated value of water intake outside the river channel is used to evaluate the satisfaction rate of the water resource demand outside the river channel, and the residual flow in the river channel is used to evaluate the satisfaction rate of the ecological water demand of the downstream storage space control section. The long-term data of the satisfaction rate of the water resource demand outside the river channel, the hydraulic elements of the water level and flow of the storage space control section, and the ecological water demand satisfaction rate are obtained; based on the long-term data, the maximum flood discharge of each water project, the storage rate at the end of the flood season, and the duration of the water level exceeding the guaranteed water level of the storage space control section are calculated;

[0027] S34. After the calculation of the combined pre-discharge scheduling condition set for the water projects is completed, the hydrological statistical model of the mathematical statistics module is used to quantitatively analyze the impact mechanism of water project scheduling on various scheduling objectives, including the response curve of the reduction in the maximum flood discharge of a single water project and the reduction in the flood season storage rate to its own pre-discharge drop level, the reduction in the water resource demand satisfaction rate outside the downstream river channel, the reduction in the ecological water demand satisfaction rate of the storage space control section, and the reduction in the over-guaranteed duration of the control section, the response curve to the pre-discharge drop level of a single water project, and the response surface to the pre-discharge drop level combination of water projects;

[0028] S35. Based on the impact mechanism of water project scheduling on various scheduling objectives, water projects are selected for joint scheduling based on the flood control, water resources, and water ecological scheduling objectives: For flood control scheduling objectives, water projects whose maximum flood discharge reduction response curves are ranked in the top 75% of all projects in descending order, and whose storage space control section over-guaranteed duration reduction response curves are ranked in the top 75% of all projects in descending order, are selected for joint scheduling; For water resources scheduling objectives, water projects whose end-of-flood storage rate reduction response curves are ranked in the bottom 75% of all projects in descending order, and whose downstream off-river water resource demand satisfaction rate reduction response curves are ranked in the bottom 75% of all projects in descending order, are selected for joint scheduling; For water ecological scheduling objectives, water projects whose end-of-flood storage rate reduction response curves are ranked in the bottom 75% of all projects in descending order, and whose storage space control section ecological water demand satisfaction rate reduction response curves are ranked in the bottom 75% of all projects in descending order, are selected for joint scheduling;

[0029] S36. Water projects selected for flood control, water resources, and water ecology scheduling objectives will be included in the scope of joint scheduling;

[0030] S37. Based on the impact mechanism of water project scheduling on various scheduling targets, and according to the allowable thresholds for the satisfaction rate of water resource demand outside the downstream river channel and the ecological water demand satisfaction rate of the storage space control section, establish the overall requirements for the joint scheduling of water projects included in the joint scheduling scope, including the thresholds for the pre-discharge drop in water level of individual water projects and the thresholds for the total pre-discharge drop in water level of water projects included in the joint scheduling scope;

[0031] S38. Industry experts refer to the overall requirements for joint water project scheduling and operation distilled from the basin data base and issue structured water project scheduling instructions based on decision-making consultations;

[0032] S39. Use the large language model of the artificial intelligence module to carry out semantic recognition of dispatch instructions and distill them into structured information, including the participating water projects, dispatch measures, and dispatch constraints. Combined with the overall requirements for the joint dispatch of water projects, a joint dispatch operation mode for water projects is formed.

[0033] S310. The water projects involved in the expert dispatching instructions shall be dispatched in accordance with the corresponding dispatching measures and dispatching constraints. Other water projects included in the joint dispatching scope shall be dispatched in accordance with their own dispatching regulations. The dispatching measures include no pre-discharge to focus on protecting water resources and water ecology, and specifying the maximum pre-discharge reduction level to focus on flood control. The dispatching constraints include the maximum flood discharge of the water project not exceeding the threshold, the total pre-discharge reduction level of each water project not exceeding the threshold, and the hydraulic elements of the downstream control section not exceeding the threshold range.

[0034] Preferably, the step S34 comprises the following steps:

[0035] S34.1. Quantitatively analyze the impact of water project scheduling on flood control objectives: Extract the pre-discharge reduction amplitude of each water project in the pre-discharge scheduling condition set, extract long-series data on the discharge flow and upstream water level of each water project, and extract long-series data on hydraulic elements of the downstream storage space control section; based on the univariate nonlinear statistics of the hydrological statistical model, construct a response curve for each water project's maximum flood discharge reduction amplitude to its own pre-discharge reduction amplitude for each water project; based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface for multiple projects based on the excess duration reduction amplitude of the storage space control section to the combination of water project pre-discharge reduction amplitudes; based on the main effect analysis of the hydrological statistical model, derive the response curve for the excess duration reduction amplitude of the storage space control section to the pre-discharge reduction amplitude of each water project;

[0036] S34.2. Quantitatively analyze the impact of water project scheduling on water resource targets: Extract the magnitude of pre-discharge reduction in each water project within the pre-discharge scheduling scenario, extract long-term data on the storage capacity of each water project, and extract long-term data on the satisfaction rate of downstream off-channel water resource demand. Based on the univariate nonlinear statistics of the hydrological statistical model, construct a response curve for each water project's end-of-flood storage rate reduction to its own pre-discharge reduction level. Based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface for multiple projects' reduction in the satisfaction rate of downstream off-channel water resource demand to the combination of water project pre-discharge reduction levels. Based on the main effect analysis of the hydrological statistical model, derive the response curve of the reduction in the satisfaction rate of downstream off-channel water resource demand to the pre-discharge reduction level of each water project.

[0037] S34.3. Quantitatively analyze the impact mechanism of water project scheduling on ecological goals: extract the amplitude of pre-discharge reduction of each water project in the pre-discharge scheduling condition set, and extract the long series data of ecological water demand satisfaction rate of the control section of the downstream storage space; based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface of the reduction amplitude of the ecological water demand satisfaction rate of the control section to the combination of water project pre-discharge reduction levels for multiple projects; based on the main effect analysis of the hydrological statistical model, obtain the response curve of the reduction amplitude of the ecological water demand satisfaction rate of the control section to the pre-discharge reduction level of a single water project.

[0038] Preferably, step S4 comprises the following steps:

[0039] S41. Based on the archive of measured extreme flood events in the basin and the archive of measured events with different occurrence frequencies, obtain the scheduling boundary conditions and treatment sections for flood events with a return period of 1,000 years, 100 years, 50 years, 20 years, and 10 years, as well as extreme flood events. For each type of flood event, generate a set of joint optimized scheduling plans for water projects, which specifically includes the following steps:

[0040] S411. Set decision variables, analysis objects, scheduling goals, and constraints:

[0041] Set decision variables, analysis objects, and scheduling objectives, determine constraints based on the joint scheduling operation mode of water projects in the basin, and build a multi-objective mathematical model for joint scheduling of water projects:

[0042] The decision variables are the calculated values ​​of the pre-discharge water level of each water regulation project and are initialized to zero;

[0043] The analysis targets the flood season water levels of water projects included in the joint dispatching scope. For the participating water projects with a specified maximum pre-discharge drop level, the flood season water level is deducted from the calculated pre-discharge drop level. For other water projects, the flood season water level is used.

[0044] The dispatching objectives include: ① the total magnitude of water level reduction by pre-discharge at each water project is as small as possible; ② the maximum change in the maximum value of hydraulic elements at the treatment section compared to the scenario where no pre-discharge is performed at the water project;

[0045] The constraints include: ① the calculated pre-discharge drop value of the participating water projects does not exceed the maximum pre-discharge drop level specified in the expert dispatching instructions; ② the maximum flood discharge of the water project does not exceed the threshold; ③ the total pre-discharge drop level of each water project does not exceed the threshold; ④ the hydraulic elements of the downstream disposal section do not exceed the threshold range;

[0046] S412. Conduct simulation calculations for water projects and downstream river channels: The flood season water levels of water projects included in the joint dispatching scope are analyzed. Combined with the flood process vectors of the water projects in the basin and the downstream sections of the flood event, simulation calculations for water projects and downstream river channels are conducted. Continuous calculations are performed using the water project control application model to obtain the discharge flow, upstream water level, and storage capacity of each water project. Combined with the basin hydrological model or hydrodynamic model of the basin simulation module, flow calculations are performed for the hydraulic elements of the downstream storage space of the water project to obtain the hydraulic elements of the control sections of the storage space, including the disposal sections.

[0047] S413, Scheduling Target and Constraint Evaluation: Calculate the values ​​of each scheduling target; evaluate whether each constraint condition is satisfied; if not, calculate the constraint violation value; otherwise, the constraint violation value is zero; sum to obtain the total constraint violation value;

[0048] S414, iterative optimization of decision variables: based on the values ​​of the decision variables and their corresponding scheduling target values ​​and the total constraint violation amount, the multi-objective heuristic algorithm of the artificial intelligence module is used to iteratively update the values ​​of the decision variables;

[0049] S415, repeatedly calling steps S412 to S414 to re-evaluate the decision variables; exiting the iteration when the number of repeated calls meets the upper limit of the iterative optimization number;

[0050] S416. When exiting the iteration, the water project joint optimization scheduling plan set is returned, including the multi-objective frontier of the decision variables, the corresponding upstream water level process of the water project, the hydraulic element process of the storage space control section including the treatment section, the numerical values ​​of each scheduling target, and the total constraint violation amount. For each plan in the water project joint optimization scheduling plan set, for the water projects included in the joint scheduling scope, the water project scheduling measures, i.e., the extent of water level reduction by pre-discharge at each water project, are identified; and the treatment effect, i.e., the extent of change in the maximum value of the hydraulic element of the storage space treatment section compared to the scenario where no pre-discharge occurs at all.

[0051] S42. Create simulated combined operating condition data and generate a water project joint optimization scheduling plan set based on the scheduling boundary conditions and treatment sections of each operating condition, specifically including the following steps:

[0052] S421. Based on the different values ​​of hydraulic elements at the basin hydrological stations and the different spatial dimensions of flood frequencies in the basin water projects and downstream areas, Latin hypercube sampling of the hydrological statistical model is used to obtain a variety of different dispatching boundary conditions; basin treatment sections are determined based on the permutation and combination method; and combinations of dispatching boundary conditions and treatment sections are enumerated to generate a variety of different flood simulation combination operating condition data.

[0053] S422. For each flood simulation combination condition, call steps S411 to S416 to obtain a set of joint optimization scheduling plans for water projects, including the multi-objective frontier of decision variables, the corresponding upstream water level process of the water project, the hydraulic element process of the storage space control section including the treatment section, the numerical values ​​of each scheduling target, and the total constraint violation amount. For each plan in the set of joint optimization scheduling plans for water projects included in the joint scheduling scope, identify the water project scheduling measures, i.e., the extent of water level reduction by pre-discharge at each water project, and identify the treatment effectiveness, i.e., the extent of change in the maximum value of the hydraulic element at the treatment section of the storage space compared to a scenario in which no pre-discharge occurs at the water project.

[0054] S43. Based on the archives of measured extreme events of floods in the river basin, archives of measured events with different occurrence frequencies, and simulated combined operating condition data, all the joint optimization scheduling plan sets for water projects are screened for each combination of scheduling boundary conditions and treatment sections. For the plan sets under any combination, the basic mathematical method of one-to-one bidirectional mapping of mathematical statistics models is used to derive the bidirectional mapping relationship between the scheduling measures of water projects included in the joint scheduling scope and the driving feedback mechanism of the treatment results.

[0055] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0056] (1) Multi-objective coverage: Existing technologies mostly focus on a single objective such as flood control. The system and method of the present invention take into account multiple objectives such as flood control, water resources, and water ecology, and can better meet the needs of comprehensive management of river basin water projects.

[0057] (2) Strong integration: Through the integration of the knowledge graph distiller and the professional model set, the system and method of the present invention realize a unified framework of data, models and expert experience knowledge, surpassing the single function of the existing technology.

[0058] (3) Real-time optimization capability: The pre-plan reasoning engine and similar scenario matching function of the present invention significantly improve the adaptability and efficiency of real-time scheduling, especially under complex watershed conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a structural diagram of the watershed water project joint intelligent scheduling system for smart water conservancy of the present invention.

[0060] Figure 2This is a flow chart of the knowledge graph distiller application for the joint scheduling of watershed water projects of the present invention.

[0061] Figure 3 This is a schematic diagram of the watershed water conservancy model set of the present invention.

[0062] Figure 4 The present invention provides a flow chart for constructing and applying the water project joint scheduling plan reasoner.

[0063] Figure 5 This is an example of the driving feedback mapping relationship between water project scheduling measures and treatment results in an embodiment of the present invention.

[0064] Figure 6 This is a diagram showing the application effect of the basin water project joint intelligent scheduling system in an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The present invention will be further described below.

[0066] Basin Background: The XJ River Basin covers approximately 3,000 square kilometers, spanning multiple administrative regions and boasting a complex terrain encompassing mountains, hills, and plains. Annual precipitation in the basin ranges from 1,000 to 1,600 mm. Precipitation is concentrated during the high-water season (April to August), accounting for over 70% of the annual water volume. Rapid ups and downs in the river are prone to flooding. During the low-water season (October to February), precipitation is scarce, resulting in a sharp drop in river flow and exposed riverbeds, leading to increased pressure on water supply and deteriorating water quality. The basin has an average annual flow of approximately 130 cubic meters per second, with an annual runoff of 4.1 billion cubic meters. The upper reaches of the basin are home to three large reservoirs (D1, D2, and D3) and two medium-sized reservoirs (Z1 and Z2). The large reservoirs have a flood control capacity exceeding 100 million cubic meters, while the medium-sized reservoirs have a combined storage capacity exceeding 10 million cubic meters. These reservoirs alleviate flood pressure downstream and ensure water supply for both downstream and downstream waterways. The downstream control section and water project operation and disposal section of the XJ River Basin is Station G. When the April 2023 flood in the XJ River Basin began, D1, D2, D3, and Z1 were all at flood control levels of 124.2, 154.4, 155.0, and 163.3 m, respectively. For the April 2023 flood, without considering the pre-discharge drawdown of individual reservoirs, the forecast results showed that the initial water level at Station G was 44.57 m, with a predicted maximum water level of 47.1 m, exceeding the guaranteed water level of 47.0 m for this control section. Furthermore, forecasts for flood process vectors, peak discharge, peak occurrence time, and flood volume are available for the catchment areas and downstream sections of the five reservoirs.

[0067] like Figures 1 to 6 As shown, the present invention is a watershed water project joint intelligent scheduling system for smart water conservancy, including a watershed water project joint scheduling knowledge graph distiller, a watershed water conservancy model set and a water project joint scheduling plan reasoner;

[0068] The knowledge graph distiller for joint scheduling of water projects in the basin, based on the basin data base and expert instructions, analyzes the impact of water project scheduling on various scheduling objectives and constructs the operation mode of joint scheduling of water projects in the basin through bidirectional distillation. Bidirectional distillation uses a large language model to analyze expert instructions (knowledge → data), and then uses statistical models to extract data patterns (data → knowledge), ultimately forming intelligent scheduling rules for human-machine collaboration.

[0069] The watershed water conservancy model set includes an engineering scheduling module, a watershed simulation module, a mathematical statistics module and an artificial intelligence module;

[0070] The project scheduling module includes a water project control application model for continuously calculating the upstream water level, storage capacity and discharge flow of the water project using the water balance principle;

[0071] Watershed simulation module: used to simulate and analyze the hydrological and hydrodynamic processes under the influence of water conservancy projects in the watershed, including watershed hydrological models and hydrodynamic models;

[0072] Mathematical Statistics Module: used to perform statistical analysis and regularity quantification on the upstream water level, storage capacity, and discharge flow of water conservancy-related river basin water projects, as well as the water level and flow at river basin hydrological stations, including hydrological statistical models and similarity analysis models;

[0073] Artificial Intelligence Module: This module is used to efficiently solve water project scheduling-related problems and make intelligent decisions using intelligent algorithms, including neural network models, multi-objective heuristic algorithms, and large language models.

[0074] The proposed joint water project scheduling plan reasoner constructs a bidirectional mapping relationship between water project scheduling measures and treatment results based on archives of measured extreme flood events in the basin, archives of measured events with different occurrence frequencies, and simulated combined operating condition data. This bidirectional mapping relationship is then reversely applied to the current treatment event to optimize the selection of participating water projects and reason about matching scheduling measures with them. This bidirectional mapping relationship transforms scheduling measures and treatment results into reversible functions through mathematical modeling, realizing an intelligent feedback mechanism that "inputs targets and outputs measures."

[0075] A scheduling method for a watershed water project joint intelligent scheduling system for smart water conservancy, comprising the following steps:

[0076] S1, collect XJ basin data base;

[0077] The XJ Basin data base includes real-time monitoring data on water levels and flow related to flood control, water resources, and water ecological scheduling objectives for water projects in the basin, compiled hydrological data at hydrological stations and reservoirs, and water resources bulletins. The hydrological stations involved include Station G, and the reservoirs involved include Stations D1, D2, D3, Z1, and Z2. The compiled hydrological data are valid until December 31, 2022, starting from January 1, 2000 (or the year following reservoir construction). The water resources bulletins are collected from 2000 to 2022.

[0078] The XJ Basin data base includes the design reports and operation regulations of the five reservoirs D1, D2, D3, Z1, and Z2, as well as information such as the defense water level, warning water level, guaranteed water level, and safe river discharge of Station G in the basin control section.

[0079] Among them, the XJ basin data base includes data archives corresponding to measured extreme flood events in the basin and measured events with different occurrence frequencies; measured extreme flood events in the basin include the "1998.6", "2008.6", "2017.7" and "2022.6" major flood events; measured events with different occurrence frequencies include flood events with a return period of one thousand years, one hundred years, fifty years, twenty years and ten years in the basin considering the composition of regional floods (reservoir catchment area, reservoir downstream area); the data archives of each flood event include scheduling boundary conditions and disposal sections, among which the scheduling boundary conditions are the initial values ​​of water levels and flows at hydrological stations in the basin before water project scheduling, as well as flood characteristics such as flood process vectors, peak flows, peak occurrence times and flood volumes of each reservoir and downstream area in the basin; the disposal sections of each flood event data archive are all G stations.

[0080] S2. Obtain instructions from experts on joint dispatch of water projects in the basin;

[0081] Among them, the expert instructions for the joint dispatch of river basin water projects include structured information such as participating water projects, dispatching measures, and dispatching constraints. Among them, when the river basin water project joint intelligent dispatching system is applied before the flood season in 2024, the participating water projects will be determined as D1, D2, and D3, and other water projects included in the joint dispatching scope will be dispatched in accordance with the approved dispatching procedures. In terms of dispatching measures, the maximum pre-discharge reduction level of D1, D2, and D3 is specified to focus on ensuring flood control in the river basin. The maximum reduction level of the three compared with their flood limit water levels (124.2, 154.4, and 155.0 m, respectively) is 1.0, 0.5, and 0.5 m. The dispatching constraints include that the maximum flood discharge of each participating water project does not exceed its own dam safety discharge capacity, the total pre-discharge reduction level of each reservoir does not exceed 1.5 m, and the peak water level of the downstream control section G station does not exceed its guaranteed water level of 47.0 m.

[0082] S3. Knowledge Graph Distiller for Joint Scheduling of River Basin Water Projects: Based on the river basin data base and the river basin hydraulic model set, this tool quantifies the impact of water project scheduling on flood control, water resources, and water ecological scheduling objectives, and establishes overall requirements for the joint scheduling of water projects included in the joint scheduling scope. It also conducts semantic recognition of expert instructions for joint scheduling of river basin water projects, including participating water projects, scheduling measures, and scheduling constraints, and formulates a joint scheduling operation mode for water projects based on the overall requirements for joint scheduling of water projects.

[0083] The step S3 comprises the following steps:

[0084] S31. Extract long-term inflow data from the XJ River Basin data base for the D1, D2, D3, Z1, and Z2 catchments and downstream sections; water resource demand and water withdrawal ratio data for production, domestic use, ecology, and agriculture outside the river channel; ecological water demand data for station G within the river channel; and water replenishment loss coefficients for water projects; dispatching procedures, characteristic parameters, and curves for each water project, including allowable water level fluctuations, water level-volume curves, water level-discharge capacity curves, and water level-controlled discharge curves; and the warning water level at station G in the downstream storage space control section of the water project is 46.0 m, and the guaranteed water level is 47.0 m.

[0085] S32. Discretize the maximum possible amplitude of 2.0 m of the pre-discharge drop compared to the flood limit water level during the D1, D2, D3, Z1, and Z2 flood seasons. Enumerate the combinations of drop amplitudes across multiple projects to construct a set of joint pre-discharge scheduling conditions for water projects. Simulate the water projects and downstream river channels for each condition in the set.

[0086] S33. Carry out simulation calculations of the water project and downstream river channel for the joint pre-discharge scheduling conditions of each water project: Use the water project control application model to carry out long-term continuous calculations to obtain long-term data on the discharge flow, upstream water level, and storage capacity of each water project. Combined with the river confluence hydrological model of the XJ basin simulation module, carry out water level and flow routing of the downstream storage space of the water project. The water flow routing takes into account the water replenishment loss of the water project, the incoming water along the section, the water intake and return of water outside the river channel. The calculated value of water intake outside the river channel is used to evaluate the satisfaction rate of water resource demand outside the river channel, and the residual flow in the river channel is used to evaluate the satisfaction rate of ecological water demand at station G of the downstream storage space control section. The satisfaction rate of water resource demand outside the river channel, the water level and flow at station G of the storage space control section, and the ecological water demand satisfaction rate are obtained in a long series. Based on the long series data, the maximum flood discharge of each water project, the storage rate at the end of the flood season, and the duration of the water level exceeding the guaranteed water level at station G of the storage space control section are calculated.

[0087] After the calculation of the combined pre-discharge scheduling condition set of S34 and water projects is completed, the hydrological statistical model of the mathematical statistics module is used to quantitatively analyze the impact mechanism of the scheduling of the D1, D2, D3, Z1, and Z2 water projects on each scheduling target, including the response curve of the reduction in the maximum flood discharge of a single water project and the reduction in the flood season storage rate to its own pre-discharge drop level, the reduction in the water resource demand satisfaction rate outside the downstream river channel, the reduction in the ecological water demand satisfaction rate of the storage space control section G station, and the reduction in the over-guaranteed duration of the control section G station, the response curve to the pre-discharge drop level of a single water project, and the response surface to the combination of pre-discharge drop levels of water projects;

[0088] The step S34 comprises the following steps:

[0089] S34.1. Quantitatively analyze the impact of water project scheduling on flood control objectives: extract the pre-discharge reduction amplitude of each water project in the pre-discharge scheduling condition set, extract long-term data on the discharge flow and upstream water level of each water project, and extract long-term data on the water level and flow at station G of the downstream storage space control section; based on the univariate nonlinear statistics of the hydrological statistical model, construct a response curve for each water project's maximum flood discharge reduction amplitude to its own pre-discharge reduction amplitude for each water project; based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface for the combination of the water project pre-discharge reduction amplitude of the over-guaranteed duration at station G of the storage space control section to the water project pre-discharge reduction amplitude for five reservoirs; based on the main effect analysis of the hydrological statistical model, derive the response curve of the over-guaranteed duration reduction amplitude at station G of the storage space control section to the pre-discharge reduction amplitude of each water project;

[0090] S34.2. Quantitatively analyze the impact of water project scheduling on water resource targets: Extract the magnitude of pre-discharge reduction in each water project in the pre-discharge scheduling scenario, extract long-term data on the storage capacity of each water project, and extract long-term data on the satisfaction rate of downstream off-channel water resources demand. Based on the univariate nonlinear statistics of the hydrological statistical model, construct a response curve for each water project's end-of-flood storage rate reduction to its own pre-discharge reduction level. Based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface for the reduction in the satisfaction rate of downstream off-channel water resources demand to the combination of water project pre-discharge reduction levels for the five reservoirs. Based on the main effect analysis of the hydrological statistical model, derive the response curve of the reduction in the satisfaction rate of downstream off-channel water resources demand to the pre-discharge reduction level of each water project.

[0091] S34.3. Quantitatively analyze the impact of water project scheduling on ecological objectives: Extract the magnitude of pre-discharge reductions for each water project in the pre-discharge scheduling scenario, and extract a long series of data on the ecological water demand satisfaction rate at station G in the downstream storage space control section. Based on multivariate nonlinear statistics of the hydrological statistical model, construct a response surface for the five reservoirs to the reduction in the ecological water demand satisfaction rate at station G in the control section to the combination of pre-discharge reduction levels of the water projects. Based on the main effect analysis of the hydrological statistical model, derive the response curve of the reduction in the ecological water demand satisfaction rate at station G in the control section to the pre-discharge reduction level of each water project.

[0092] S35. Based on the impact mechanism of water project scheduling on various scheduling objectives, water projects are selected for joint scheduling based on the flood control, water resources, and water ecological scheduling objectives: For flood control scheduling objectives, water projects whose maximum flood discharge reduction response curves are ranked in the top 75% of all projects in descending order, and whose over-guaranteed duration reduction response curves at station G of the storage space control section are ranked in the top 75% of all projects in descending order, are selected for joint scheduling; For water resources scheduling objectives, water projects whose end-of-flood storage rate reduction response curves are ranked in the bottom 75% of all projects in descending order, and whose downstream off-river water resource demand satisfaction rate reduction response curves are ranked in the bottom 75% of all projects in descending order, are selected for joint scheduling; For water ecological scheduling objectives, water projects whose end-of-flood storage rate reduction response curves are ranked in the bottom 75% of all projects in descending order, and whose ecological water demand satisfaction rate reduction response curves at station G of the storage space control section are ranked in the bottom 75% of all projects in descending order, are selected for joint scheduling;

[0093] S36. The four reservoirs D1, D2, D3, and Z1, which were jointly selected for flood control, water resources, and water ecology scheduling goals, will be included in the scope of joint scheduling;

[0094] S37. Based on the impact mechanism of water project scheduling on various scheduling targets, and in accordance with the 5% threshold for the satisfaction rate of downstream off-channel water resource demand and the ecological water demand satisfaction rate at station G in the storage space control section, establish overall requirements for joint scheduling operations for water projects included in the joint scheduling scope, including the threshold for the pre-discharge drop in water level for individual water projects and the threshold for the total pre-discharge drop in water level for water projects included in the joint scheduling scope;

[0095] S38. When the basin water project joint intelligent dispatching system is applied before the 2024 flood season, industry experts will refer to the overall requirements for joint dispatching of water projects distilled from the basin data base and issue structured water project dispatching instructions based on decision-making consultations;

[0096] S39. Use the large language model of the artificial intelligence module to carry out semantic recognition of dispatch instructions and distill them into structured information, including the participating water projects, dispatch measures, and dispatch constraints. Combined with the overall requirements for the joint dispatch of water projects, a joint dispatch operation mode for water projects is formed.

[0097] S310. The water projects involved in the expert dispatching instructions shall be dispatched in accordance with the corresponding dispatching measures and dispatching constraints. Other water projects included in the joint dispatching scope shall be dispatched in accordance with their own dispatching regulations. The dispatching measures include no pre-discharge to focus on protecting water resources and water ecology, and specifying the maximum pre-discharge reduction level to focus on flood control. The dispatching constraints include the maximum flood discharge of the water project not exceeding the threshold, the total pre-discharge reduction level of each water project not exceeding the threshold, and the hydraulic elements of the downstream control section not exceeding the threshold range. Among them, the participating water projects are identified as D1, D2, and D3; the maximum pre-discharge water level reduction amplitudes of D1, D2, and D3 are specified to focus on ensuring flood control in the basin. The maximum water level reduction amplitudes of the three compared with their flood limit water levels (124.2, 155.0, and 154.4m respectively) are 1.0, 0.5, and 0.5 m respectively; other water projects included in the joint dispatching scope are dispatched in accordance with the approved dispatching regulations; the dispatching constraints include that the maximum flood discharge of each participating water project does not exceed its own dam safety discharge volume, the total pre-discharge water level reduction amplitude of each reservoir does not exceed 1.5 m, and the flood peak water level of the downstream control section G station does not exceed the guaranteed water level by 1.0 m, that is, 48.0 m.

[0098] S4. Constructing a water project joint dispatch plan reasoner: Based on the XJ basin data base's flood extreme event archives, archives of events with different frequencies, and simulated combined operating condition data, and according to the basin water project joint dispatch operation mode and basin hydraulic model set, a set of water project joint optimized dispatch plans is generated, establishing a bidirectional mapping relationship between water project dispatch measures and the driving feedback mechanism of treatment results;

[0099] The step S4 comprises the following steps:

[0100] S41. Based on the XJ River Basin flood extreme event archive and the archive of events with different frequencies, obtain the scheduling boundary conditions and treatment sections for flood events with a return period of 1,000 years (P = 0.1%), 1,000 years (P = 1%), 1,50 years (P = 2%), 1,20 years (P = 5%), and 1,00 years (P = 10%), as well as extreme flood events. For each type of flood event, generate a set of joint optimization scheduling plans for water projects, which specifically include the following steps:

[0101] S411. Set decision variables, analysis objects, scheduling goals, and constraints:

[0102] Set decision variables, analysis objects, and scheduling objectives, determine constraints based on the joint scheduling operation mode of water projects in the basin, and build a multi-objective mathematical model for joint scheduling of water projects:

[0103] The decision variables are the pre-discharge drop level trial values ​​of each water diversion project D1, D2, and D3, and are initialized to 0.0 m;

[0104] The analysis targets the flood season water levels of water projects D1, D2, D3, and Z1 included in the joint dispatch scope. For the participating water projects D1, D2, and D3, which have specified maximum pre-discharge reduction levels, the flood season water level is deducted from the calculated pre-discharge reduction level. For other water projects, the flood season water level is used.

[0105] The dispatching objectives include: ① the total amplitude of the pre-discharge water level reduction of D1, D2, and D3 is as small as possible; ② the change of the flood peak water level at station G of the treatment section is the largest compared with the scenario where the water project does not pre-discharge at all;

[0106] The constraints include: ① the calculated pre-discharge drop values ​​for participating water projects do not exceed the maximum pre-discharge drop ranges specified in the expert dispatching instructions (1.0, 0.5, and 0.5 m); ② the maximum flood discharge of each water project does not exceed the threshold (safe discharge capacity below the dam); ③ the total pre-discharge drop range of each water project does not exceed the 1.5 m threshold; and ④ the hydraulic elements of the downstream disposal section do not exceed the threshold range (the peak water level at station G does not exceed 48.0 m).

[0107] S412. Conduct simulation calculations for water projects and downstream river channels: The flood season water levels of water projects D1, D2, D3, and Z1 included in the joint dispatch scope are analyzed. Combined with the flood process vectors of the water projects in the basin and the downstream intervals of the flood event, simulation calculations for water projects and downstream river channels are conducted. Continuous calculations are performed using the control and application models of the D1, D2, D3, and Z1 water projects to obtain the discharge flow, upstream water level, and storage capacity of each water project. Combined with the river confluence hydrological model of the watershed simulation module, water level and flow calculations are performed for the downstream storage space of the water project to obtain the water level and flow process at station G, the control section of the storage space.

[0108] S413, Scheduling Target and Constraint Evaluation: Calculate the values ​​of each scheduling target; evaluate whether each constraint condition is satisfied; if not, calculate the constraint violation value; otherwise, the constraint violation value is zero; sum to obtain the total constraint violation value;

[0109] S414, iterative optimization of decision variables: Based on the values ​​of the decision variables and their corresponding scheduling target values ​​and the total constraint violation amount, the multi-objective heuristic algorithm Non-dominated Sorting Genetic Algorithm II (NSGA-II) of the artificial intelligence module is used to iteratively update the values ​​of the decision variables;

[0110] S415, repeatedly calling steps S412 to S414 to re-evaluate the decision variables; exiting the iteration when the number of repeated calls meets the upper limit of the iterative optimization number;

[0111] S416. When exiting the iteration, the water project joint optimization scheduling plan set is returned, including the multi-objective Pareto frontier of the decision variables, the corresponding upstream water level processes of the D1, D2, D3, and Z1 water projects, the water level and flow processes of the G station in the storage space control section, the values ​​of each scheduling target, and the total constraint violation amount; for each plan in the water project joint optimization scheduling plan set, for the water projects included in the joint scheduling scope, the water project scheduling measures, i.e., the amplitude of the pre-discharge reduction of the water level of each water project, are marked, and the treatment results, i.e., the amplitude of the change of the decision variables of the water level at the G station in the storage space treatment section compared to the pre-discharge reduction level, are marked.

[0112] S42. Create simulated combined operating condition data and generate a water project joint optimization scheduling plan set based on the scheduling boundary conditions and treatment sections of each operating condition, specifically including the following steps:

[0113] S421. Considering the complex spatial dimensions of different water levels or flows at hydrological stations in the basin, and the different frequencies of flooding in water projects and downstream areas, Latin hypercube sampling of the hydrological statistical model is used to obtain a variety of different dispatching boundary conditions. The basin treatment section is determined based on the permutation and combination method. Given that the treatment section in the XJ basin is station G, the only treatment section option is station G. The dispatching boundary conditions and treatment section combinations are enumerated to generate a variety of different flood simulation combined operating condition data.

[0114] S422. For each flood simulation combination condition, call steps S411 to S416 to obtain a set of joint optimization scheduling plans for water projects, including the multi-objective Pareto front of decision variables, the corresponding upstream water level processes of the D1, D2, D3, and Z1 water projects, the water level and flow processes of station G at the storage space control section, the values ​​of each scheduling target, and the total constraint violation amount; for each plan in the set of joint optimization scheduling plans for water projects included in the joint scheduling scope, identify the water project scheduling measures, i.e., the extent of water level reduction by pre-discharge at each water project, and identify the treatment effectiveness, i.e., the extent of change in the water level at station G at the storage space treatment section compared to a scenario in which no pre-discharge occurs at all;

[0115] S43. Based on the archives of measured extreme events of flood in the basin, archives of measured events with different occurrence frequencies, and simulated combined operating condition data, all the joint optimization scheduling plans for all water projects are screened for each combination of scheduling boundary conditions and treatment sections. For the plan set under any combination, the basic mathematical method of one-to-one correspondence and bidirectional mapping of mathematical statistics models is used to derive the bidirectional mapping relationship between the scheduling measures of water projects included in the joint scheduling scope and the driving feedback mechanism of the treatment results. The mapping relationship diagram is shown in Figure 5 ;

[0116] S5. During the 2023 flood season, the user uses the water project joint scheduling plan reasoner to handle the "2023.4" flood event in the XJ River Basin. The user conducts reasoning on the optimization of participating water projects and the matching of scheduling measures. The user obtains the scheduling boundary conditions for the "2023.4" flood event (see the basin background section for details), the treatment section (station G), and the user-defined treatment effect. The user-defined treatment effect is that the maximum flood level at station G drops by 0.2 m, that is, drops to 0.1 m below the guaranteed water level.

[0117] Using the similarity analysis model of the mathematical statistics module, similar flood scenarios are matched with the current event scheduling boundary conditions and treatment sections in the measured extreme event archives, measured event archives with different occurrence frequencies and simulated combined operating condition data; based on the two-way mapping relationship between the water project scheduling measures and the treatment results of similar scenarios, the two-way mapping relationship of the driving feedback mechanism is reversely applied according to the treatment results set by the user to match the pre-discharge drop level amplitude of the water projects included in the joint scheduling scope, among which the water projects with pre-discharge drop level amplitude greater than zero are regarded as the preferred results of the participating water projects, and the corresponding pre-discharge drop level amplitude is regarded as the scheduling measure matching reasoning result.

[0118] The water project joint dispatch plan reasoner outputs the optimization results of the participating water projects and the matching reasoning results of the dispatch measures. That is, for the "2023.4 flood", only a single water project D2 will be pre-discharged, with a pre-discharge lowering of the water level by 0.35 m; the other water projects included in the joint dispatch scope (D1, D3, Z1) will maintain flood limits and be dispatched according to the dispatch regulations.

[0119] After the user executed the output of the water project joint dispatching plan inference engine, the flood level at station G in the XJ basin control section increased during the flood season due to the pre-discharge of the upstream D2 reservoir, but the measured peak water level dropped by about 0.2 m compared with the forecast value (see Figure 6 ), successfully dropped below the guaranteed water level of station G.

[0120] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A watershed water project joint intelligent dispatching system for smart water conservancy, characterized by: It includes a knowledge graph distiller for joint scheduling of water projects in a river basin, a water conservancy model set in a river basin, and a reasoner for joint scheduling plans of water projects; The knowledge graph distiller for joint scheduling of water projects in the basin, based on the basin data base and expert instructions, analyzes the impact mechanism of water project scheduling on various scheduling targets, and constructs the operation mode of joint scheduling of water projects in the basin through bidirectional distillation; The watershed water conservancy model set includes an engineering scheduling module, a watershed simulation module, a mathematical statistics module, and an artificial intelligence module. It applies water conservancy mathematical models, statistical models, and artificial intelligence models to serve the knowledge graph distillation and water project joint scheduling plan reasoning of the watershed water project joint scheduling. The water project joint scheduling plan reasoner, based on the measured extreme event archives of flood in the basin, the measured event archives of different occurrence frequencies, and the simulated combined operating condition data, constructs a bidirectional mapping relationship between the driving feedback mechanism of water project scheduling measures and treatment results. When responding to the current treatment event, the bidirectional mapping relationship of the driving feedback mechanism is reversely applied to carry out the optimization of participating water projects and the matching reasoning of participating water project scheduling measures. Carry out the reasoning of the optimization of the participating water projects and the matching of the dispatching measures: obtain the dispatching boundary conditions of the current disposal event, the disposal section selected by the user and the disposal effect set by the user, and use the similarity analysis model of the mathematical statistics module to match the similar flood scenarios with the dispatching boundary conditions and the disposal section of the current disposal event in the measured extreme event archives, the measured event archives with different occurrence frequencies and the simulated combined working condition data. Based on the bidirectional mapping relationship of the driving feedback mechanism between the water project dispatching measures and the disposal effect of similar scenarios, the bidirectional mapping relationship of the driving feedback mechanism is reversely applied according to the user-set disposal effect to match the pre-discharge drop level amplitude of the water projects included in the joint dispatching scope. Among them, the water projects with the pre-discharge drop level amplitude greater than zero are regarded as the optimization results of the participating water projects, and the corresponding pre-discharge drop level amplitude is regarded as the dispatching measure matching reasoning result. The water project joint dispatching plan reasoner outputs the optimization results of the participating water projects and the dispatching measure matching reasoning results.

2. The watershed water project joint intelligent dispatching system for smart water conservancy according to claim 1 is characterized by: The basin data base is real-time monitoring data, compiled data, statistical bulletins, water project design reports and scheduling regulations, and characteristic water levels of basin control sections related to flood control, water resources, and water ecological scheduling goals of basin water projects. It contains data archives corresponding to measured extreme events of basin floods and measured events with different occurrence frequencies. The data archives include scheduling boundary conditions and disposal sections. The scheduling boundary conditions are the initial values ​​of water levels and flows at basin hydrological stations before water project scheduling, as well as flood process vectors, peak flows, peak occurrence times, and flood volumes of basin water projects and downstream intervals. The expert instructions of the water project joint scheduling knowledge graph distiller are instructions from water conservancy industry experts for the joint scheduling of water projects, including participating water projects, scheduling measures, and scheduling constraints.

3. The watershed water project joint intelligent dispatching system for smart water conservancy according to claim 1 is characterized by: The engineering scheduling module includes a water engineering control application model; The water project control application model applies the water balance principle to carry out continuous calculations at a preset time step during the scheduling period to obtain the upstream water level, storage capacity and discharge flow process of the water project; at each calculation step, relying on the water level-volume curve, water level-discharge capacity curve and water level-controlled discharge curve of the water project, the real-time water storage capacity, upper and lower limits of the discharge capacity of the water project are analyzed, and the water level and discharge flow requirements in the water project scheduling regulations are faced. The period discharge flow is calculated in combination with the period inflow flow, and the upstream water level and storage capacity at the end of the water project period are deduced based on the water balance.

4. A dispatching method for a watershed water project joint intelligent dispatching system for smart water conservancy according to claims 1 to 3, characterized in that: The following steps are involved: S1. Collect the data base of a certain watershed; S2. Obtain instructions from experts on joint dispatch of water projects in the basin; S3. Knowledge Graph Distiller for Joint Scheduling of River Basin Water Projects: Based on the river basin data base and the river basin hydraulic model set, this tool quantifies the impact of water project scheduling on flood control, water resources, and water ecological scheduling objectives, and establishes overall requirements for the joint scheduling of water projects included in the joint scheduling scope. It also conducts semantic recognition of expert instructions for joint scheduling of river basin water projects, including participating water projects, scheduling measures, and scheduling constraints, and formulates a joint scheduling operation mode for water projects based on the overall requirements for joint scheduling of water projects. S4. Constructing a water project joint dispatch plan reasoner: Based on the flood extreme event archives, the archives of events with different frequencies, and the simulated combined operating condition data in the basin data base, and according to the basin water project joint dispatch operation mode and basin hydraulic model set, a set of water project joint optimized dispatch plans is generated. This establishes a bidirectional mapping relationship between water project dispatch measures and the driving feedback mechanism of the treatment results. S5. Use the water project joint scheduling plan reasoner to respond to the current disposal event and carry out the optimization of the participating water projects and the matching reasoning of the scheduling measures: obtain the scheduling boundary conditions of the current disposal event, the user-selected disposal section and the user-set disposal effect, and use the similarity analysis model of the mathematical statistics module to match the similar flood scenarios with the scheduling boundary conditions and disposal sections of the current disposal event in the measured extreme event archives, the measured event archives with different occurrence frequencies and the simulated combined working condition data. Based on the bidirectional mapping relationship between the water project scheduling measures and the disposal effect of the similar scenarios, the bidirectional mapping relationship of the driving feedback mechanism is reversely applied according to the user-set disposal effect to match the pre-discharge drop level amplitude of the water projects included in the joint scheduling scope, among which the water projects with the pre-discharge drop level amplitude greater than zero are regarded as the optimization results of the participating water projects, and the corresponding pre-discharge drop level amplitude is regarded as the scheduling measure matching reasoning result. The water project joint scheduling plan reasoner outputs the optimization results of the participating water projects and the scheduling measure matching reasoning results.

5. The scheduling method of the watershed water project joint intelligent scheduling system for smart water conservancy according to claim 4 is characterized by: The step S3 comprises the following steps: S31. Extract long-term inflow data for each water project's catchment area and downstream sections from the basin data base, water resource demand and water withdrawal ratio data for production, domestic, ecological, and agricultural water use outside the river channel, ecological water demand data at different locations within the river channel, and water project replenishment loss coefficients; extract the scheduling procedures, characteristic parameters, and curves for each water project, including allowable water level fluctuations, water level-volume curves, water level-discharge capacity curves, and water level-controlled discharge curves; and extract the warning water level and guaranteed water level for the control section of the downstream storage space of each water project. S32. Discretize the maximum possible amplitude of the pre-discharge water level drop of each water project compared to the flood limit water level during the flood season, enumerate the combinations of water level drop amplitudes among multiple water projects, construct a water project joint pre-discharge scheduling operating condition set, and perform simulation calculations for the water project and downstream river channel for each operating condition in the operating condition set; S33. Carry out simulation calculations of the water project and downstream river channel for the joint pre-discharge scheduling conditions of each water project: Use the water project control application model to carry out long-term continuous calculations to obtain long-term data on the discharge flow, upstream water level, and storage capacity of each water project. Combined with the basin hydrological model or hydrodynamic model of the basin simulation module, carry out water flow routing of the hydraulic elements of the downstream storage space of the water project. The water flow routing takes into account the water replenishment loss of the water project, the incoming water along the section, the water intake and return of water outside the river channel. The calculated value of water intake outside the river channel is used to evaluate the satisfaction rate of the water resource demand outside the river channel, and the residual flow in the river channel is used to evaluate the satisfaction rate of the ecological water demand of the downstream storage space control section. The long-term data of the satisfaction rate of the water resource demand outside the river channel, the hydraulic elements of the water level and flow of the storage space control section, and the ecological water demand satisfaction rate are obtained; based on the long-term data, the maximum flood discharge of each water project, the storage rate at the end of the flood season, and the duration of the water level exceeding the guaranteed water level of the storage space control section are calculated; S34. After the calculation of the combined pre-discharge scheduling condition set for water projects is completed, the hydrological statistical model of the mathematical statistics module is used to quantitatively analyze the impact mechanism of water project scheduling on various scheduling targets, including the response curve of the reduction in the maximum flood discharge of a single water project and the reduction in the flood season storage rate to its own pre-discharge drop level; the reduction in the water resource demand satisfaction rate outside the downstream river channel, the reduction in the ecological water demand satisfaction rate of the storage space control section, and the reduction in the over-guaranteed duration of the control section; the response curve of the pre-discharge drop level of a single water project and the response surface of the pre-discharge drop level combination of water projects; S35. Based on the impact mechanism of water project scheduling on various scheduling objectives, water projects are selected for joint scheduling based on the flood control, water resources, and water ecological scheduling objectives: For flood control scheduling objectives, water projects whose maximum flood discharge reduction response curves are ranked in the top 75% of all projects in descending order, and whose storage space control section over-guaranteed duration reduction response curves are ranked in the top 75% of all projects in descending order, are selected for joint scheduling; For water resources scheduling objectives, water projects whose end-of-flood storage rate reduction response curves are ranked in the bottom 75% of all projects in descending order, and whose downstream off-river water resource demand satisfaction rate reduction response curves are ranked in the bottom 75% of all projects in descending order, are selected for joint scheduling; For water ecological scheduling objectives, water projects whose end-of-flood storage rate reduction response curves are ranked in the bottom 75% of all projects in descending order, and whose storage space control section ecological water demand satisfaction rate reduction response curves are ranked in the bottom 75% of all projects in descending order, are selected for joint scheduling; S36. Water projects selected for flood control, water resources, and water ecology scheduling objectives will be included in the scope of joint scheduling; S37. Based on the impact mechanism of water project scheduling on various scheduling targets, and according to the allowable thresholds for the satisfaction rate of water resource demand outside the downstream river channel and the ecological water demand satisfaction rate of the storage space control section, establish the overall requirements for the joint scheduling of water projects included in the joint scheduling scope, including the thresholds for the pre-discharge drop in water level of individual water projects and the thresholds for the total pre-discharge drop in water level of water projects included in the joint scheduling scope; S38. Industry experts refer to the overall requirements for joint water project scheduling and operation distilled from the basin data base and issue structured water project scheduling instructions based on decision-making consultations; S39. Use the large language model of the artificial intelligence module to carry out semantic recognition of dispatch instructions and distill them into structured information, including the participating water projects, dispatch measures, and dispatch constraints. Combined with the overall requirements for the joint dispatch of water projects, a joint dispatch operation mode for water projects is formed. S310. The water projects involved in the expert dispatching instructions shall be dispatched in accordance with the corresponding dispatching measures and dispatching constraints. Other water projects included in the joint dispatching scope shall be dispatched in accordance with their own dispatching regulations. The dispatching measures include no pre-discharge to focus on protecting water resources and water ecology or specifying the maximum pre-discharge reduction level to focus on flood control. The dispatching constraints include the maximum flood discharge of the water project not exceeding the threshold, the total pre-discharge reduction level of each water project not exceeding the threshold, and the hydraulic elements of the downstream control section not exceeding the threshold.

6. The dispatching method of the watershed water project joint intelligent dispatching system of the smart water conservancy according to claim 5 is characterized by: The step S34 comprises the following steps: S34.

1. Quantitatively analyze the impact of water project scheduling on flood control objectives: Extract the pre-discharge reduction amplitude of each water project in the pre-discharge scheduling condition set, extract long-series data on the discharge flow and upstream water level of each water project, and extract long-series data on hydraulic elements of the downstream storage space control section; based on the univariate nonlinear statistics of the hydrological statistical model, construct a response curve for each water project's maximum flood discharge reduction amplitude to its own pre-discharge reduction amplitude for each water project; based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface for multiple projects based on the excess duration reduction amplitude of the storage space control section to the combination of water project pre-discharge reduction amplitudes; based on the main effect analysis of the hydrological statistical model, derive the response curve for the excess duration reduction amplitude of the storage space control section to the pre-discharge reduction amplitude of each water project; S34.

2. Quantitatively analyze the impact of water project scheduling on water resource targets: Extract the magnitude of pre-discharge reduction in each water project within the pre-discharge scheduling scenario, extract long-term data on the storage capacity of each water project, and extract long-term data on the satisfaction rate of downstream off-channel water resource demand. Based on the univariate nonlinear statistics of the hydrological statistical model, construct a response curve for each water project's end-of-flood storage rate reduction to its own pre-discharge reduction level. Based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface for multiple projects' reduction in the satisfaction rate of downstream off-channel water resource demand to the combination of water project pre-discharge reduction levels. Based on the main effect analysis of the hydrological statistical model, derive the response curve of the reduction in the satisfaction rate of downstream off-channel water resource demand to the pre-discharge reduction level of each water project. S34.

3. Quantitatively analyze the impact mechanism of water project scheduling on ecological goals: extract the amplitude of pre-discharge reduction of each water project in the pre-discharge scheduling condition set, and extract the long series data of ecological water demand satisfaction rate of the control section of the downstream storage space; based on the multivariate nonlinear statistics of the hydrological statistical model, construct a response surface of the reduction amplitude of the ecological water demand satisfaction rate of the control section to the combination of water project pre-discharge reduction levels for multiple projects; based on the main effect analysis of the hydrological statistical model, obtain the response curve of the reduction amplitude of the ecological water demand satisfaction rate of the control section to the pre-discharge reduction level of a single water project.

7. The scheduling method of the watershed water project joint intelligent scheduling system for smart water conservancy according to claim 4 is characterized by: The step S4 comprises the following steps: S41. Based on the archive of measured extreme flood events in the basin and the archive of measured events with different occurrence frequencies, obtain the scheduling boundary conditions and treatment sections for flood events with a return period of 1,000 years, 100 years, 50 years, 20 years, and 10 years, as well as extreme flood events. For each type of flood event, generate a set of joint optimized scheduling plans for water projects, which specifically includes the following steps: S411. Set decision variables, analysis objects, scheduling goals, and constraints: Set decision variables, analysis objects, and scheduling objectives, determine constraints based on the joint scheduling operation mode of water projects in the basin, and build a multi-objective mathematical model for joint scheduling of water projects: The decision variables are the calculated values ​​of the pre-discharge water level of each water regulation project and are initialized to zero; The analysis targets the flood season water levels of water projects included in the joint dispatching scope. For the participating water projects with a specified maximum pre-discharge drop level, the flood season water level is deducted from the calculated pre-discharge drop level. For other water projects, the flood season water level is used. The dispatching objectives include: ① the total magnitude of water level reduction by pre-discharge at each water project is as small as possible; ② the maximum change in the maximum value of hydraulic elements at the treatment section compared to the scenario where no pre-discharge is performed at the water project; The constraints include: ① the calculated pre-discharge drop value of the participating water projects does not exceed the maximum pre-discharge drop level specified in the expert dispatching instructions; ② the maximum flood discharge of the water project does not exceed the threshold; ③ the total pre-discharge drop level of each water project does not exceed the threshold; ④ the hydraulic elements of the downstream disposal section do not exceed the threshold range; S412. Conduct simulation calculations for water projects and downstream river channels: The flood season water levels of water projects included in the joint dispatching scope are analyzed. Combined with the flood process vectors of the water projects in the basin and the downstream sections of the flood event, simulation calculations for water projects and downstream river channels are conducted. Continuous calculations are performed using the water project control application model to obtain the discharge flow, upstream water level, and storage capacity of each water project. Combined with the basin hydrological model or hydrodynamic model of the basin simulation module, flow calculations are performed for the hydraulic elements of the downstream storage space of the water project to obtain the hydraulic elements of the control sections of the storage space, including the disposal sections. S413, Scheduling Target and Constraint Evaluation: Calculate the values ​​of each scheduling target; evaluate whether each constraint condition is satisfied; if not, calculate the constraint violation value; otherwise, the constraint violation value is zero; The sum is obtained to get the total constraint violation; S414, iterative optimization of decision variables: based on the values ​​of the decision variables and their corresponding scheduling target values ​​and the total constraint violation amount, the multi-objective heuristic algorithm of the artificial intelligence module is used to iteratively update the values ​​of the decision variables; S415, repeatedly calling steps S412 to S414 to re-evaluate the decision variables; exiting the iteration when the number of repeated calls meets the upper limit of the iterative optimization number; S416. When exiting the iteration, the water project joint optimization scheduling plan set is returned, including the multi-objective frontier of the decision variables, the corresponding upstream water level process of the water project, the hydraulic element process of the storage space control section including the treatment section, the numerical values ​​of each scheduling target, and the total constraint violation amount. For each plan in the water project joint optimization scheduling plan set, for the water projects included in the joint scheduling scope, the water project scheduling measures, i.e., the extent of water level reduction by pre-discharge at each water project, are identified; and the treatment effect, i.e., the extent of change in the maximum value of the hydraulic element of the storage space treatment section compared to the scenario where no pre-discharge occurs at all. S42. Create simulated combined operating condition data and generate a water project joint optimization scheduling plan set based on the scheduling boundary conditions and treatment sections of each operating condition, specifically including the following steps: S421. Based on the different values ​​of hydraulic elements at the basin hydrological stations and the different spatial dimensions of flood frequencies in the basin water projects and downstream areas, Latin hypercube sampling of the hydrological statistical model is used to obtain a variety of different dispatching boundary conditions; basin treatment sections are determined based on the permutation and combination method; and combinations of dispatching boundary conditions and treatment sections are enumerated to generate a variety of different flood simulation combination operating condition data. S422. For each flood simulation combination condition, call steps S411 to S416 to obtain a set of joint optimization scheduling plans for water projects, including the multi-objective frontier of decision variables, the corresponding upstream water level process of the water project, the hydraulic element process of the storage space control section including the treatment section, the numerical values ​​of each scheduling target, and the total constraint violation amount. For each plan in the set of joint optimization scheduling plans for water projects included in the joint scheduling scope, identify the water project scheduling measures, i.e., the extent of water level reduction by pre-discharge at each water project, and identify the treatment effectiveness, i.e., the extent of change in the maximum value of the hydraulic element at the treatment section of the storage space compared to a scenario in which no pre-discharge occurs at the water project. S43. Based on the archives of measured extreme events of floods in the river basin, archives of measured events with different occurrence frequencies, and simulated combined operating condition data, all the joint optimization scheduling plan sets for water projects are screened for each combination of scheduling boundary conditions and treatment sections. For the plan sets under any combination, the basic mathematical method of one-to-one bidirectional mapping of mathematical statistics models is used to derive the bidirectional mapping relationship between the scheduling measures of water projects included in the joint scheduling scope and the driving feedback mechanism of the treatment results.

Images