Drainage basin water resource intelligent allocation system and method based on dynamic graph data model

Abstract

The invention discloses a dynamic graph data model-based basin water resource intelligent allocation system and method, and the method comprises the steps: constructing a space-time coupled basin dynamic graph model (ST-DWGM), carrying out the fusion modeling of satellite remote sensing data, an IoT sensor network and water right transaction information, employing a hypergraph structure to represent a multi-dimensional constraint relation, and carrying out the calculation of the multi-dimensional constraint relation. And a quantum evolutionary algorithm is introduced to realize minute-level multi-objective optimization solution. A weather forecast event is responded through a dynamic topology reconstruction mechanism, a non-dominated solution set is generated in combination with Pareto frontier analysis, and finally an intelligent decision-making scheme considering the ecological flow red line, the water delivery energy consumption cost and the water supply guarantee rate is formed. Compared with a traditional method, the method has the remarkable advantages in the capacity of dynamic response, complex relation modeling, cross-regional water resource coordination and the like, and an innovative solution is provided for intelligent allocation of drainage basin water resources.

Description

Technical Field

[0001] The present invention belongs to the field of smart water conservancy technology, and in particular relates to a system and method for intelligent allocation of water resources in a river basin based on a dynamic graph data model. Background Art

[0002] River basin water resources management is a crucial foundation for ensuring regional water security, maintaining ecological balance, and supporting sustainable economic and social development. With the intensification of global climate change and the increase in the intensity of human activities, river basin systems are facing multiple challenges, including altered runoff patterns, imbalanced temporal and spatial distribution of water resources, and increased water pollution. Traditional management models are no longer able to address the complex water-energy-food coupling relationship, particularly in cross-administrative coordinated scheduling scenarios where the need to protect ecological baseflows is paramount while balancing industrial and agricultural water demands and energy consumption constraints.

[0003] The existing technical system has three key flaws: at the data modeling level, static topological models still rely on historical hydrological data for construction, and it is difficult to integrate dynamic data sources such as meteorological satellite forecasts and real-time monitoring by IoT sensors, resulting in serious delays in scheduling plans during sudden rainfall or droughts. For example, during the flood peak in the Pearl River Basin in 2023, the traditional model did not integrate radar precipitation forecast data, and the scheduling response was delayed by 6 hours; at the relationship expression level, there is a lack of quantitative modeling tools for the complex coupling relationship between multiple water sources (surface runoff, groundwater, recycled water) and multiple users (agricultural irrigation areas, industrial parks, and ecological protection areas). In particular, cross-administrative water rights transactions have an actual transaction success rate of less than 45% due to the inefficiency of the supply and demand matching algorithm; at the optimization mechanism level, the current single-objective optimization method cannot coordinate conflicting goals such as ecological flow guarantee, water transmission cost control, and water supply stability. The 2024 Yellow River Basin scheduling test showed that the traditional linear programming scheme would cause a 62% surge in water transmission energy consumption when meeting 95% of the ecological flow. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a basin water resources intelligent allocation system and method based on a dynamic graph data model. The dynamic graph model based on time-space coupling integrates multi-source dynamic data, bids farewell to traditional static dependence, and can quickly adjust the scheduling plan in the face of sudden water conditions, thereby improving the response speed.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: A system and method for intelligent allocation of water resources in a river basin based on a dynamic graph data model, comprising the following steps: A dynamic graph model construction module is used to construct a spatiotemporally coupled watershed dynamic graph model, including physical nodes, virtual control nodes, and dynamic edge weights. The physical nodes include hydrological stations, reservoirs, and water intakes, and the virtual control nodes are dynamically generated through demand forecasting and weather forecasting. A topology reconstruction module is used to drive the expansion or contraction of network topology based on rainfall intensity or drought index events, adding flood diversion channel edges or pruning inefficient water transfer paths; A multi-dimensional constraint hypergraph module connects ecological, economic, and social constraint nodes through hyperedges to quantify ecological base flow protection, energy consumption budget control, and water supply reliability goals; Multi-source data fusion module, which uses Kalman filtering to fuse satellite remote sensing and ground sensing data, and predicts future hydrological conditions based on LSTM neural network; A quantum evolutionary optimization engine, which transforms the deployment problem into a multi-objective path optimization and generates a Pareto frontier solution set through quantum coding, adaptive mutation, and non-dominated sorting; The real-time decision module selects the optimal scheduling solution from the solution set based on fuzzy comprehensive evaluation and outputs control instructions.

[0006] Preferably, the update formula of the dynamic edge weight is: ; in, For real-time traffic, is the maximum flow rate, is the pipeline efficiency coefficient (0.7-0.95), Compensation for facility health, Indicates the health status score of the facility.

[0007] Preferably, the topology reconstruction module responds to rainfall events according to the following rules: when the local rainfall intensity is monitored When the terrain gradient is based on DEM Generate flood diversion channel edge set .

[0008] Preferably, the multidimensional constraint hyperedges of the multidimensional constraint hypergraph module include: Ecological constraints beyond the boundary ,satisfy ; Economic constraint beyond the limit , constraining the total energy consumption: ; Social constraints beyond the boundary , water supply guarantee rate required: .

[0009] A method for operating a basin water resources intelligent allocation system based on a dynamic graph data model comprises the following steps: S1. Construct a dynamic graph data model for the watershed, including dynamic node networks, topology reconstruction, and multi-dimensional constrained hypergraph modeling; S2. Multi-objective optimization based on quantum evolutionary algorithms, including: Encode the deployment plan as a qubit vector and evaluate ecological, economic and social goals through fitness functions; The Lagrangian relaxation method is used to deal with the constraints, and the population is optimized by combining adaptive mutation and elite retention strategies. An improved NSGA-II framework was used to generate a non-dominated solution set, and the final scheduling scheme was selected through fuzzy comprehensive evaluation. The evaluation index system includes: ecological flow compliance rate (weight 0.35-0.5), water transmission energy consumption cost ratio (weight 0.3-0.45), and water supply guarantee rate (weight 0.2-0.25). The weights are dynamically adjusted according to the drought level, with the ecological weight increased to 0.5 during the dry season.

[0010] Preferably, the fitness function in step S2 is: ; The weight coefficient is dynamically adjusted through the hierarchical analysis method.

[0011] Preferably, each deployment scheme in the quantum coding scheme is expressed as dimensional qubit vector ,satisfy ,in and is the complex probability amplitude.

[0012] Preferably, the weight of the membership function in the fuzzy comprehensive evaluation is dynamically adjusted according to the drought level, and the ecological weight in the dry season is increased to 0.5.

[0013] Preferably, the model is updated every 15 minutes, and the event queue processing mechanism controls the computational complexity to , the update delay is less than 5 seconds at a scale of tens of thousands of nodes.

[0014] Preferably, the control instructions output by the real-time decision module include gate opening, pump station power and flood diversion channel opening and closing instructions.

[0015] The present invention can achieve the following beneficial effects: 1. A dynamic graph model based on spatiotemporal coupling integrates multi-source dynamic data, saying goodbye to traditional static dependence. In the face of sudden water conditions, the scheduling plan can be quickly adjusted to improve response speed.

[0016] 2. Integrate multi-dimensional constrained hypergraphs with multi-source data to quantify the complex relationships between multiple water sources and multiple users, optimize cross-regional water rights trading algorithms, and promote the efficient flow of water resources.

[0017] 3. Combining quantum evolution with fuzzy comprehensive evaluation breaks through the limitations of single objectives, coordinates multiple objectives such as ecology, economy, and society, and maximizes the benefits of river basin water resources management.

[0018] 4. The model is updated frequently to control computational complexity, ensure real-time data processing, output precise control instructions, and realize refined and intelligent management of water resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The present invention will be further described below with reference to the accompanying drawings and examples: Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0020] The preferred solution is Figure 1 As shown in the figure, a basin water resources intelligent allocation system based on a dynamic graph data model includes the following modules: A dynamic graph model construction module is used to construct a spatiotemporally coupled watershed dynamic graph model, including physical nodes, virtual control nodes, and dynamic edge weights. The physical nodes include hydrological stations, reservoirs, and water intakes, and the virtual control nodes are dynamically generated through demand forecasting and weather forecasting. A topology reconstruction module is used to drive the expansion or contraction of network topology based on rainfall intensity or drought index events, adding flood diversion channel edges or pruning inefficient water transfer paths; A multi-dimensional constraint hypergraph module connects ecological, economic, and social constraint nodes through hyperedges to quantify ecological base flow protection, energy consumption budget control, and water supply reliability goals; Multi-source data fusion module, which uses Kalman filtering to fuse satellite remote sensing and ground sensing data, and predicts future hydrological conditions based on LSTM neural network; A quantum evolutionary optimization engine, which transforms the deployment problem into a multi-objective path optimization and generates a Pareto frontier solution set through quantum coding, adaptive mutation, and non-dominated sorting; The real-time decision module selects the optimal scheduling solution from the solution set based on fuzzy comprehensive evaluation and outputs control instructions. The update formula of the dynamic edge weight is: ; in, For real-time traffic, is the maximum flow rate, is the pipeline efficiency coefficient (0.7-0.95), Compensation for facility health, Indicates the health status score of the facility.

[0021] The topology reconstruction module responds to rainfall events according to the following rules: when the local rainfall intensity is monitored When the terrain gradient is based on DEM Generate flood diversion channel edge set .

[0022] The multi-dimensional constraint hyperedges of the multi-dimensional constraint hypergraph module include: Ecological constraints beyond the boundary ,satisfy ; Economic constraint beyond the limit , constraining the total energy consumption: ; Social constraints beyond the boundary , water supply guarantee rate required: .

[0023] A method for operating a basin water resources intelligent allocation system based on a dynamic graph data model comprises the following steps: S1, construction of watershed dynamic graph data model; S11. Dynamic Node-Edge Network Modeling: In this paper, a dynamic node network is used to represent the physical facilities and virtual control units of the water network in the basin. In order to achieve full-factor digital mapping of water resources, we established a dual-modal node system.

[0024] Physical nodes include physical facilities such as hydrological stations (monitoring real-time water levels), reservoirs (water storage capacity constraints), and water intakes (user demand interfaces), and their spatial topology is precisely anchored through the GIS coordinate system.

[0025] Virtual control node It is dynamically generated by the formula: ; in, is the generating function, Forecasting the future through LSTM neural network The demand forecast value for a period of time, the input features include 8-dimensional parameters such as historical water consumption, temperature, holiday factors, etc. This is the rainfall intensity forecasted by the weather forecast. The data comes from Doppler radar monitoring. d is the adaptive adjustment coefficient (the default value is 1.0, which is automatically increased to 1.2 during drought warning). or 、 m is the weight coefficient (under normal conditions or =0.6, m =0.4, flood season m This type of node can adaptively adjust the proportion of water distribution across regions, for example, automatically creating flood diversion channel nodes when heavy rainfall is predicted.

[0026] Edge weight The dynamic update mechanism of is realized through a multi-parameter coupling model: ; Where, For real-time traffic, is the maximum flow rate, is the pipeline efficiency coefficient (0.7-0.95), Compensation for facility health, This represents the health status of facilities such as pump stations and gates, captured in real time by IoT sensors. This design allows water delivery capacity calculations to account for both physical constraints and the impact of equipment status.

[0027] S12. Topological reconstruction of spatiotemporal coupling To respond to dramatic changes in the hydrological system caused by sudden changes in meteorological conditions, we established event-driven topology reconstruction rules: Rainfall-driven expansion: When satellite remote sensing detects local rainfall intensity Time-triggered flooding network generation algorithm adds a new edge set : ; in, The terrain gradient calculated for the DEM digital elevation model ensures that the newly added edges conform to the natural confluence law.

[0028] Drought response contraction: based on the Joint Drought Index Dynamic evaluation (SPI is the Standard Precipitation Index, SRI is the Surface Runoff Index), when Perform topology pruning when: ; in, For in time The edge set when Represents a connected node and nodes The edge, Represents an edge water transport capacity, The maximum water delivery capacity.

[0029] This mechanism prioritizes the retention of high-priority water transmission routes, such as livelihood water supply pipelines, and implements drought emergency response through dynamic topology adjustment.

[0030] S13. Hypergraph Modeling of Multidimensional Constraints: In order to solve the ecological, economic and social multi-objective coordination problem, the present invention innovatively introduces hypergraph theory to establish a constraint network.

[0031] Convert the three types of constraints into hyperedge connection rules: Ecological constraints beyond the boundary : Connect all river section nodes that need to ensure minimum ecological flow, meeting the following requirements: ; in, For nodes The outflow rate, Minimum ecological flow, For the basin area, ensure that the ecological base flow is adapted to the basin scale.

[0032] Economic constraint beyond the limit : Aggregate cost-sensitive nodes such as water pumping stations and purification plants to constrain total energy consumption: ; in, is the power of the pumping station, is the running time, For energy budget, It is a factor of electricity price fluctuation to achieve economic dynamic budget control.

[0033] Social constraints beyond the boundary : Construct a three-level guarantee chain of "water source-water plant-user", requiring: ; in, is the water supply, is the demand, is the weight coefficient, Improve the water pressure sensor compliance rate for end users and enhance water supply reliability.

[0034] S14. Multi-source data fusion and model updating To achieve high model fidelity, we designed a hierarchical data fusion architecture.

[0035] Spatial Data Calibration: Fusion of Satellite Remote Sensing Data via Kalman Filtering and ground sensor data : ; For the calibrated spatial data, Kalman gain Dynamically adjust according to sensor accuracy to solve the problem of insufficient spatiotemporal resolution of remote sensing data.

[0036] Time series prediction module: uses LSTM neural network to predict the hydrological status in the next 15 minutes: ; in, is the hidden state of the LSTM neural network at time t, is the weight matrix from hidden state to hidden state, is the weight matrix input to the hidden state, input feature Contains 12-dimensional parameters such as rainfall and evaporation in the previous 6 hours, It is the bias term of the hidden state and outputs the predicted values ​​of water level and flow in the future period.

[0037] Dynamic update engine: perform a full map update every 15 minutes, and control the computational complexity to , ensuring that the update delay is less than 5 seconds at a scale of tens of thousands of nodes.

[0038] S2, a multi-objective path optimization method based on quantum evolution to solve the intelligent allocation of water resources in a river basin.

[0039] S21. Problem Transformation and Quantum Coding: In this paper, the water resource allocation problem on the dynamic graph is transformed into a multi-objective path optimization problem. Aiming at the complex connection relationship between nodes in the watershed dynamic graph model ST-DWGM, a coding scheme based on quantum bit superposition state is designed. Each allocation scheme is represented as dimensional qubit vector : ; in Characterizes the probability of water flow distribution, and is the complex probability amplitude. This encoding method can express multiple potential deployment schemes simultaneously, significantly improving the efficiency of exploring the solution space.

[0040] S22. Fitness Evaluation and Constraint Processing: In order to quantify the pros and cons of the allocation scheme, a fitness function F is constructed that includes the three-dimensional objectives of ecology, economy and society: ; Where, is the actual ecological flow, is the minimum ecological flow requirement, is the benchmark water delivery energy cost, is the energy cost of water transportation, represents the number of users who meet their water supply needs, is the total number of users. The weight coefficient is dynamically adjusted through the hierarchical analysis method to adapt to different scheduling scenarios.

[0041] For hypergraph constraints, Lagrangian relaxation method is used to deal with hard constraints: ; in, is the Lagrangian function, For ecological flow constraints, is the energy budget constraint, To constrain the water supply guarantee rate, is the threshold of the constraint condition, and Corresponding. By dynamically adjusting the penalty factor , achieving a balance between constraint violation and target optimization.

[0042] S23. Quantum Evolutionary Optimization: First, the population is initialized, and the initial feasible solution set is generated based on the network flow algorithm. The classical solution is converted into a quantum state through the quantum rotating gate: ; in, Indicates the The classic solution is The value of the bit gene, represents the size of the population, that is, the total number of classical solutions in the population, is the quantum rotation angle.

[0043] Then, we conduct crossover mutation operation and design a two-point controlled quantum crossover operator to retain the dominant gene fragments before and after the crossover point. : ; is the maximum number of iterations, which is a condition for the termination of the algorithm. Increase and gradually enhance local search capabilities.

[0044] The elite retention strategy uses fast non-dominated sorting and crowding calculation to select Pareto frontier solutions from the parent and child populations. The capacity of the external archive set is set to 1.5 times the population size to ensure the diversity and convergence of the solution set.

[0045] S24. Generation of multi-objective Pareto solutions: Generate non-dominated solution sets through the improved NSGA-II framework: (1) Calculate the dominance level and crowding distance of each solution; (2) Perform fast non-dominated sorting on the mixed population and divide the frontier level; (3) Select individuals in the frontier layer in descending order of crowding degree; (4) When it is necessary to Layer Selection When there are individuals, the solution with large congestion degree is retained first; The time complexity of this process is controlled in ,in is the target number, is the population size, which is significantly better than the traditional algorithm .

[0046] S25. Real-time decision-making and control execution: The optimization engine outputs the Pareto frontier solution every 15 minutes, and the final scheduling plan is selected through a fuzzy comprehensive evaluation method. This fuzzy comprehensive evaluation utilizes a three-level indicator system: the first-level indicators include ecological benefits (weighting 0.35-0.5), economic benefits (weighting 0.3-0.45), and social benefits (weighting 0.2-0.25). The weights are dynamically adjusted based on the drought severity (e.g., the ecological weight is increased to 0.5 during the dry season). Second-level indicators include ecological flow compliance rate, river ecological health index, water transmission energy consumption ratio, equipment maintenance costs, water supply guarantee rate, and user satisfaction. Third-level indicators are specific quantitative parameters (e.g., ecological flow compliance rate = actual ecological flow / target ecological flow). Basic weights are first determined using the Analytic Hierarchy Process (AHP) and then dynamically adjusted based on the drought severity (SPI index). Membership functions, such as Gaussian and trapezoidal, are used to calculate the membership of each indicator. A fuzzy matrix is ​​constructed from the standardized indicator values, and a weighted composite operation is performed to determine the comprehensive evaluation score. The plan with the highest score is ultimately selected as the optimal scheduling plan, generating control instructions for gate opening, pump station power, and other parameters.

[0047] The above embodiments are merely preferred technical solutions of the present invention and should not be construed as limiting the present invention. The scope of protection of the present invention shall be the technical solutions set forth in the claims, including equivalent alternatives to the technical features of the technical solutions set forth in the claims. In other words, equivalent alternatives and improvements within this scope are also within the scope of protection of the present invention.

Claims

1. A basin water resources intelligent allocation system based on dynamic graph data model, characterized by include: A dynamic graph model construction module is used to construct a spatiotemporally coupled watershed dynamic graph model, including physical nodes, virtual control nodes, and dynamic edge weights. The physical nodes include hydrological stations, reservoirs, and water intakes. The virtual control nodes are dynamically generated through demand forecasting and weather forecasting. The generation rule is: based on the LSTM neural network, the water demand for the next 24 hours is predicted, combined with the rainfall intensity from the weather forecast, and generated using the generating function Φ. A topology reconstruction module is used to drive the expansion or contraction of network topology based on rainfall intensity or drought index events, adding flood diversion channel edges or pruning inefficient water transfer paths; A multi-dimensional constraint hypergraph module connects ecological, economic, and social constraint nodes through hyperedges to quantify ecological base flow protection, energy consumption budget control, and water supply reliability goals; Multi-source data fusion module, which uses Kalman filtering to fuse satellite remote sensing and ground sensing data, and predicts future hydrological conditions based on LSTM neural network; A quantum evolutionary optimization engine, which transforms the deployment problem into a multi-objective path optimization and generates a Pareto frontier solution set through quantum coding, adaptive mutation, and non-dominated sorting; The real-time decision module selects the optimal scheduling solution from the solution set based on fuzzy comprehensive evaluation and outputs control instructions.

2. The intelligent water resource allocation system for a river basin based on a dynamic graph data model according to claim 1 is characterized in that: The update formula of the dynamic edge weight is: ； in, For real-time traffic, is the maximum flow rate, is the pipeline efficiency coefficient, Compensation for facility health, Indicates the health status score of the facility.

3. The intelligent water resource allocation system for a river basin based on a dynamic graph data model according to claim 1 is characterized in that: The topology reconstruction module responds to rainfall events according to the following rules: when the local rainfall intensity is monitored When the terrain gradient is based on DEM Generate flood diversion channel edge set .

4. The intelligent water resource allocation system for a river basin based on a dynamic graph data model according to claim 1 is characterized in that: The multi-dimensional constraint hyperedges of the multi-dimensional constraint hypergraph module include: Ecological constraints beyond the boundary ,satisfy ; Economic constraint beyond the limit , constraining the total energy consumption: ; Social constraints beyond the boundary , water supply guarantee rate required: ; in γ is the electricity price fluctuation factor, Δ T is the temperature change, λ is the weight coefficient and Esensor Indicates the sensor compliance rate.

5. The method for operating a watershed water resources intelligent allocation system based on a dynamic graph data model according to any one of claims 1 to 4, characterized in that The following steps are involved: S1. Construct a dynamic graph data model for the watershed, including dynamic node networks, topology reconstruction, and multi-dimensional constraint hypergraph modeling; S2. Multi-objective optimization based on quantum evolutionary algorithm, including: Encode the deployment plan as a qubit vector and evaluate ecological, economic and social goals through fitness functions; The Lagrangian relaxation method is used to deal with the constraints, and the population is optimized by combining adaptive mutation and elite retention strategies. The improved NSGA-II framework is used to generate non-dominated solution sets, and the final scheduling scheme is selected through fuzzy comprehensive evaluation. The evaluation index system includes: ecological flow compliance rate, water transmission energy consumption cost ratio, and water supply guarantee rate. The weights are dynamically adjusted according to the drought level, and the ecological weight is increased to 0.5 during the dry season.

6. The method for operating a watershed water resources intelligent allocation system based on a dynamic graph data model according to claim 5, characterized in that: The fitness function in step S2 is: ； The weight coefficient is dynamically adjusted through the hierarchical analysis method.

7. The method for operating a watershed water resources intelligent allocation system based on a dynamic graph data model according to claim 5, characterized in that: Each deployment scheme in the quantum coding scheme is expressed as dimensional qubit vector ,satisfy ,in and is the complex probability amplitude.

8. The method for operating a watershed water resources intelligent allocation system based on a dynamic graph data model according to claim 5, characterized in that: The weight of the membership function in the fuzzy comprehensive evaluation is dynamically adjusted according to the drought level, and the ecological weight in the dry season is increased to 0.

5.

9. The method for operating a basin water resources intelligent allocation system based on a dynamic graph data model according to claim 5, characterized in that: The model is updated every 15 minutes, and the event queue processing mechanism controls the computational complexity to , the update delay is less than 5 seconds at a scale of tens of thousands of nodes.

10. The method for operating a basin water resources intelligent allocation system based on a dynamic graph data model according to claim 5, characterized in that: The control instructions output by the real-time decision module include gate opening, pump station power and flood diversion channel opening and closing instructions.

Images