Reservoir scheduling decision support system for multi-objective optimization

Abstract

The invention discloses a reservoir scheduling decision support system for multi-objective optimization, and relates to the technical field of hydraulic engineering scheduling. According to the system, comprehensive optimization of flood control, power generation and ecological targets is realized through cooperative work of the multi-source dynamic constraint analysis module, the collaborative optimization module, the scheduling module and the correction module. A dynamic constraint boundary table is generated by collecting meteorological, hydrological and ecological data in real time. And matching a multi-target coordination rule through a physical constraint double-depth network algorithm, generating a Pareto feasible solution set, and mapping the Pareto feasible solution set into a coordination strategy of flood discharge flow and power generation output. In combination with a hydraulic model, flood control and power generation weights are adjusted, an executable dispatching instruction is generated, it is ensured that the ecological flow reaches the standard, and the influence of a downstream submerged area is optimized. And correcting the dynamic constraint boundary table according to actual monitoring data, and continuously optimizing a scheduling decision. Accurate management and multi-objective optimization of reservoir dispatching are realized, and the utilization efficiency of water resources is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy project scheduling, and specifically to a reservoir scheduling decision support system for multi-objective optimization. Background Technique

[0002] With the frequent occurrence of global climate change and extreme meteorological events, the reasonable scheduling and management of water resources have become increasingly complex. As an important water resource regulation facility, the scheduling decision of the reservoir directly affects multiple objectives such as flood control safety, power generation efficiency, and ecological protection. Traditional reservoir scheduling methods usually focus on single-objective optimization, such as flood control or power generation, and often ignore the coordination and balance among multiple objectives. With the diversification and comprehensiveness of the social demand for water resource management, single-objective scheduling strategies are difficult to meet the complex social, economic, and ecological environment needs. Therefore, a reservoir scheduling decision support system based on multi-objective optimization has emerged, aiming to achieve the coordinated optimization of flood control, power generation, and ecological protection objectives through advanced optimization algorithms and real-time data analysis technology, so as to improve the utilization efficiency of water resources, reduce risks, and promote sustainable development.

[0003] However, the existing reservoir scheduling decision support systems mostly rely on static mathematical models or single-objective optimization strategies, and are difficult to effectively handle the dynamic conflict problems among flood control, power generation, and ecological protection, have insufficient adaptability to high-dimensional non-linear scenarios, are prone to falling into local optima, and cannot integrate real-time meteorological, hydrological, and ecological monitoring data, resulting in the scheduling scheme lagging behind the actual working conditions. At the same time, decision-makers cannot intuitively adjust the objective weights or quantitatively evaluate the ecological risks, and cross-regional collaborative scheduling depends on manual negotiation due to insufficient transparency, with low efficiency and lack of rigid guarantee for constraint execution. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a reservoir scheduling decision support system for multi-objective optimization, which solves the problems in the above background technique.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a reservoir scheduling decision support system for multi-objective optimization, including the following modules: a multi-source dynamic constraint analysis module, a collaborative optimization module, a scheduling module, and a correction module; the multi-source dynamic constraint analysis module is used to access meteorological forecast data, hydrological monitoring data and ecological sensor data in real time through the Internet of Things, extract the competitive characteristics of flood control, power generation and ecological goals through dynamic threshold calculation, and generate a dynamic constraint boundary table with variable priority; the collaborative optimization module is used to match multi-objective collaborative rules according to the dynamic constraint boundary table through a physical constraint double-depth network algorithm to generate a Pareto feasible solution set that meets flood control safety, power generation benefits and ecological flow standards, and map the solution set into a collaborative strategy for flood discharge flow and power generation output; the scheduling module is used to dynamically adjust the flood control and power generation weights through a visual interface based on the Pareto feasible solution set, combine the hydraulic model to preview the impact path of flood discharge on the downstream flooded area and ecological flow of the reservoir, and generate executable scheduling instructions; the correction module is used to record the scheduling decision process through blockchain evidence according to the executable scheduling instructions, and correct the dynamic constraint boundary table according to the actual ecological flow monitoring data.

[0006] Furthermore, the specific process of extracting the competitive characteristics of flood control, power generation and ecological goals through dynamic threshold calculation is as follows: the lower limit of flood control reservoir capacity demand is calculated based on meteorological forecast data and hydrological monitoring data, the upper limit of power generation benefit is derived by combining the power grid load curve and the unit efficiency model, and the minimum ecological flow threshold is dynamically generated according to the dissolved oxygen and fish migration data monitored by the ecological sensor; by comparing the overlap rate of flood control reservoir capacity demand and power generation benefit interval, the conflict intensity between flood control and power generation goals is determined; when the ecological sensor monitors that the dissolved oxygen is lower than the threshold or the fish activity is abnormal, the ecological flow constraint priority is increased, and the calculation results are output as a dynamic competitive characteristic matrix of flood control, power generation and ecological goals.

[0007] Furthermore, the specific process of generating a dynamic constraint boundary table with variable priorities is as follows: based on the dynamic competitive characteristic matrix, the constraint intervals of flood control, power generation, and ecology are defined respectively, including the lower limit of flood control reservoir capacity demand, the upper limit of the feasible range of power generation output, and the rigid threshold of ecological flow; when the rainfall forecast intensity exceeds the historical threshold for the same period, the priority of the lower limit of flood control reservoir capacity is increased and the upper limit interval of power generation output is compressed; when the ecological sensor detects abnormal dissolved oxygen or fish activity, the ecological flow threshold is set as an insurmountable rigid constraint and the upper limit of power generation output is simultaneously lowered; if there is no overlap between the flood control and power generation constraint intervals, the upper limit of power generation output is recalculated with flood control as the highest priority; the adjusted constraint intervals and priority labels are encoded into a dynamic constraint boundary table.

[0008] Furthermore, according to the dynamic constraint boundary table, the specific process of matching the multi-objective cooperation rules through the physical constraint double-depth network algorithm to generate a Pareto feasible solution set that meets flood control safety, power generation benefits, and ecological flow standards is as follows: Define the action space based on the lower limit of flood control storage capacity demand, the upper limit of the feasible range of power generation output, and the rigid threshold of ecological flow in the dynamic constraint boundary table; Embed the hydrodynamic equation as a physical constraint into the network state transition rule to limit the flood discharge flow not to exceed the safe capacity of the downstream river channel; Design a multi-objective reward function, including a penalty term for flood control water level control error, a gain term for power generation benefits, and a reward term for the ecological flow compliance rate; Optimize the network parameters through offline historical data pre-training and online real-time data fine-tuning, and output a Pareto feasible solution set that meets the dynamic constraints of flood control safety, power generation benefits, and ecological flow.

[0009] Furthermore, the specific process of mapping the solution set to the cooperation strategy of flood discharge flow and power generation output is as follows: Analyze the flood discharge flow, power generation output, and ecological water release instructions in the Pareto feasible solution set, and extract the maximum flood discharge flow, the peak power generation output during the period, and the ecological flow compliance period; Based on the current reservoir gate state and unit operation conditions, map the flood discharge flow to the gate opening control instructions including flow, corresponding opening, and response time through a pre-trained hydrodynamic lookup table model; Combine the real-time load demand of the power grid and the unit ramp rate limit to generate a power generation output curve for each time period; Monitor the downstream flow in real time according to the rigid threshold of ecological flow. If the actual value is lower than the threshold, dynamically increase the opening of the water release gate according to the shortage ratio and simultaneously reduce the upper limit of power generation output; Encode the flood discharge gate opening, power generation output curve, and ecological water release instructions into a cooperation control instruction sequence according to the time stamp, and push it to the reservoir execution terminal through the industrial communication protocol.

[0010] Furthermore, based on the Pareto feasible solution set, the specific process of dynamically adjusting the flood control and power generation weights through the visualization interface is as follows: Load the lower limit of flood control storage capacity demand, the upper limit of power generation benefits, and the rigid threshold of ecological flow in the dynamic constraint boundary table into the visualization interface; Adjust the flood control weight and power generation weight through drag-and-drop slider interaction to generate a weight allocation ratio; Recalculate the Pareto front based on the adjusted weights and screen the solution subset that meets the weight interval; Generate a heat map of flood control risk levels in real time.

[0011] Furthermore, the specific process of generating executable dispatching instructions is as follows: the hydraulic model is driven according to the hydrological monitoring data, the rigid threshold of ecological flow and the coordinated control instruction sequence to simulate the propagation path and inundation range of the flood discharge flow, predict the maximum water depth and duration of the inundation area downstream of the reservoir, and evaluate the dissolved oxygen recovery time and the fish habitat compliance rate to see whether they meet the dynamic constraint boundary table; if not, the Pareto solution set is regenerated and the coordinated control instruction sequence is updated; if satisfied, the verified coordinated control instruction sequence is pushed to the execution terminal according to the timestamp encoding.

[0012] Furthermore, the specific process of correcting the dynamic constraint boundary table according to the actual ecological flow monitoring data is as follows: the actual ecological flow data of the downstream of the reservoir is collected in real time through ecological sensors. If the continuous set time monitoring value is lower than the ecological flow rigid threshold defined in the dynamic constraint boundary table, it is judged that the ecological constraint is not met, and the ecological flow constraint priority is increased, and the lower limit of flood control storage capacity demand and the upper limit of the feasible range of power generation output are compressed, triggering the regeneration of the Pareto feasible solution set and the update of the collaborative control instruction sequence.

[0013] The present invention has the following beneficial effects:

[0014] (1) The reservoir operation decision support system for multi-objective optimization integrates meteorological forecast, hydrological monitoring and ecological sensor data in real time through the multi-source dynamic constraint analysis module, extracts multi-objective competitive characteristics based on dynamic threshold calculation, and generates a dynamic constraint boundary table with variable priority, which significantly improves the data-driven decision-making ability in complex environments and ensures the dynamic priority adaptation of flood control, power generation and ecological goals. The collaborative optimization module adopts a physical constraint dual-deep network algorithm to embed the hydrodynamic equations into multi-objective collaborative rules, generate Pareto feasible solution sets that conform to physical laws and dynamic constraints, and map them into collaborative strategies for flood discharge and power generation, breaking through the adaptability limitations of traditional algorithms to nonlinear problems, ensuring the engineering feasibility and multi-objective balance of the solution set, and effectively improving the safety and economy of the scheduling scheme.

[0015] (2) A reservoir scheduling decision support system for multi-objective optimization dynamically adjusts the target weights through the visual human-computer interaction interface of the scheduling module, and combines a high-precision hydraulic model to simulate the impact path of flood discharge on downstream inundation and ecological flow, thereby achieving an intuitive quantitative assessment of risk-benefit and enhancing decision transparency and scheme acceptability. The correction module uses blockchain technology to store scheduling instructions and decision logic in an unalterable manner, and dynamically corrects the constraint boundary table based on real-time feedback data from ecological sensors, forming a "decision-making-execution-feedback-optimization" closed loop, ensuring trusted coordination of cross-regional scheduling and rigid execution of dynamic constraints, and providing full-link traceability technical support for sustainable water resources management.

[0016] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of the reservoir scheduling decision support system for multi-objective optimization of the present invention. DETAILED DESCRIPTION

[0018] The embodiment of the present application solves the problems of insufficient real-time data fusion, poor engineering feasibility of scheduling schemes and low credibility of cross-regional collaboration under dynamic conflicts among flood control, power generation and ecological protection goals through a reservoir scheduling decision support system oriented to multi-objective optimization.

[0019] The overall idea of the solution in the embodiments of this application is as follows:

[0020] Through the Internet of Things, real-time access to meteorological forecast data, hydrological monitoring data and ecological sensor data is achieved, and the competitive characteristics of flood control, power generation and ecological goals are extracted through dynamic threshold calculation to generate a dynamic constraint boundary table with variable priorities.

[0021] According to the dynamic constraint boundary table, the multi-objective coordination rules are matched through the physical constraint dual-depth network algorithm to generate a Pareto feasible solution set that meets the standards of flood control safety, power generation benefits and ecological flow, and map the solution set into a coordination strategy of flood discharge and power generation output.

[0022] Based on the Pareto feasible solution set, the weights of flood control and power generation are dynamically adjusted through a visual interface, and the impact path of flood discharge on the flooded area downstream of the reservoir and ecological flow is simulated in combination with the hydraulic model to generate executable scheduling instructions.

[0023] According to the executable scheduling instructions, the scheduling decision process is recorded through blockchain evidence storage, and the dynamic constraint boundary table is corrected according to the actual ecological flow monitoring data.

[0024] See also Figure 1, the embodiment of the present invention provides a technical solution: a reservoir scheduling decision support system for multi-objective optimization, including the following modules: a multi-source dynamic constraint analysis module, a collaborative optimization module, a scheduling module, and a correction module; the multi-source dynamic constraint analysis module is used to access meteorological forecast data, hydrological monitoring data and ecological sensor data in real time through the Internet of Things, extract the competitive characteristics of flood control, power generation and ecological goals through dynamic threshold calculation, and generate a dynamic constraint boundary table with variable priority; the collaborative optimization module is used to match multi-objective collaborative rules according to the dynamic constraint boundary table through a physical constraint double-depth network algorithm to generate a Pareto feasible solution set that meets flood control safety, power generation benefits and ecological flow standards, and map the solution set into a collaborative strategy of flood discharge flow and power generation output; the scheduling module is used to dynamically adjust the flood control and power generation weights through a visual interface based on the Pareto feasible solution set, combine the hydraulic model to preview the impact path of flood discharge on the downstream flooded area and ecological flow of the reservoir, and generate executable scheduling instructions; the correction module is used to record the scheduling decision process through blockchain evidence according to the executable scheduling instructions, and correct the dynamic constraint boundary table according to the actual ecological flow monitoring data.

[0025] In this implementation plan, the multi-source dynamic constraint analysis module: adopts a dynamic threshold calculation algorithm to analyze the competitive characteristics among multiple objectives (such as the storage capacity conflict between flood control and power generation), and generates a dynamic constraint boundary table with variable priorities. For example, when heavy rainfall is predicted, the priority of the flood control storage capacity demand is increased, and the feasible interval for power generation is compressed. Dynamic threshold calculation: Dynamically calculates the key thresholds of each objective (such as the lower limit of flood control storage capacity demand, the upper limit of power generation benefit, and the rigid threshold of ecological flow) based on real-time data. Dynamic constraint boundary table: A matrix rule table containing the variable priorities and feasible intervals of flood control, power generation, and ecological objectives (such as compressing the feasible interval for power generation when the flood control weight is increased). Integrates multi-source data in real time to dynamically generate a constraint rule table with variable priorities, providing dynamic boundary conditions for subsequent optimization to ensure that the scheduling strategy adapts to the real-time environment. The collaborative optimization module: Based on the dynamic constraint boundary table, generates a Pareto solution set that meets multi-objective collaboration, and transforms the abstract solution set into an executable flood discharge and power generation collaboration strategy for the project (such as gate opening instructions, unit output curves), solving the problem of insufficient adaptability of traditional algorithms in non-linear scenarios. Physical constraint double-depth network algorithm: A deep reinforcement learning algorithm that combines physical models (such as hydrodynamic equations), restricts the action space by embedding physical rules (such as flood discharge flow ≤ downstream safety capacity), and avoids generating invalid solutions that violate natural laws. Pareto feasible solution set: The set of all optimal solutions that cannot be further optimized among multiple conflicting objectives (flood control, power generation, ecology), and each solution represents a trade-off plan. The deduction and verification module: Allows decision-makers to adjust the objective weights according to actual needs, and verifies the feasibility of different plans (such as inundation risk, ecological restoration period) through hydraulic model deduction, generating the final executable instructions to improve decision-making transparency and plan acceptability. Visual interface dynamically adjusts weights: Real-time adjusts the priority ratio of flood control and power generation objectives (such as 70% for flood control and 30% for power generation) through a graphical interaction tool (such as dragging a slider). Hydraulic model deduction: Based on a high-precision hydraulic model (such as HEC-RAS), simulates the propagation path of flood discharge flow, and predicts the downstream inundation range and ecological restoration effect. The collaborative evidence storage module: Achieves non-tamperable evidence storage of scheduling instructions and cross-regional collaborative trusted execution, and at the same time dynamically corrects constraint conditions (such as increasing ecological priority) through a closed-loop feedback mechanism to ensure continuous optimization of the system. Blockchain evidence storage: Uses blockchain technology to record key data such as scheduling instructions and constraint correction logic to ensure that the data is non-tamperable and traceable. Dynamic constraint correction: Automatically adjusts the priorities and feasible intervals in the dynamic constraint boundary table according to the actual flow data feedback by ecological sensors (such as the dissolved oxygen concentration continuously being lower than the threshold).

[0026] Specifically, the specific process of extracting the competitive characteristics of flood control, power generation, and ecological objectives through dynamic threshold calculation is as follows: Calculate the lower limit of flood control storage capacity demand based on meteorological forecast data and hydrological monitoring data, deduce the upper limit of power generation benefits by combining the grid load curve and the unit efficiency model, and at the same time dynamically generate the minimum ecological flow threshold according to the dissolved oxygen and fish migration data monitored by ecological sensors; Determine the conflict intensity between flood control and power generation objectives by comparing the overlap rate between the flood control storage capacity demand and the power generation benefit interval. When the ecological sensor detects that the dissolved oxygen is lower than the threshold or the fish activity is abnormal, raise the priority of the ecological flow constraint, and output the calculation result as the dynamic competitive characteristic matrix of flood control, power generation, and ecological objectives.

[0027] In this implementation plan, the dynamic threshold calculation process, step 1: Calculation of the lower limit of flood control storage capacity demand: Input data: Meteorological forecast data: Forecast of basin rainfall in the next T hours; Hydrological monitoring data: Current reservoir water level, inflow, and downstream river channel safety capacity. The formula is as follows: ; ; Parameters: : Predicted total flood volume; : Runoff coefficient; : Basin catchment area; : Predicted rainfall period; : Available storage capacity of the current reservoir; : Lower limit of flood control storage capacity demand. Calculation logic: Predict the total flood volume through the rainfall-runoff model , after deducting the current storage capacity, the flood control storage capacity to be reserved is obtained. If the predicted value is less than the current storage capacity, no reservation is required. Step 2: Derivation of the upper limit of power generation efficiency Input data: Grid load curve: power demand in different time periods (such as 80MW demand in peak period and 50MW in valley period); Unit efficiency model: Turbine efficiency-output curve (such as the highest efficiency when the output is 50MW). Calculation logic: Combine load demand and unit efficiency to derive the upper limit of power generation efficiency (for example: the maximum output does not exceed 60MW to avoid efficiency decline). Step 3: Ecological flow threshold generation Input data: Ecological sensor data: dissolved oxygen concentration, fish migration activity monitoring value (such as spawning flow demand); Ecological response model: relationship between dissolved oxygen recovery rate and flow, fish habitat integrity index. Calculation logic: Dynamically generate the minimum ecological flow threshold (for example: the flow needs to be maintained ≥30m³ / s and dissolved oxygen ≥5mg / L during the spawning period). Step 4: Competitive feature matrix generation: Conflict intensity determination: Compare the lower limit of flood control storage capacity demand (such as 120 million m³) with the available storage capacity corresponding to the upper limit of power generation efficiency (such as power generation requires the release of part of the storage capacity), and calculate the overlap rate of the two (i.e. the storage capacity interval that can be satisfied together). Example: If the flood control demand occupies 80% of the storage capacity and the power generation demand occupies 60% of the storage capacity, and the overlap rate is low (such as only 40%), it is determined that the conflict intensity between flood control and power generation goals is high. Ecological priority adjustment: When the ecological sensor detects that the dissolved oxygen concentration is lower than the threshold (such as 5 mg / L) or the fish activity is abnormal (such as spawning is blocked), the ecological flow constraint priority is raised to the highest. Example: If the dissolved oxygen concentration is lower than the threshold, the flow rate is mandatory to be ≥30 m³ / s, and it is executed before flood control and power generation needs. Dynamic feature matrix output: The dynamic values of flood control storage capacity demand, power generation efficiency upper limit, and ecological flow threshold (such as storage capacity of 120 million m³, output of 60MW, flow of 30m³ / s) and conflict intensity (such as high / medium / low) are integrated into a matrix for real-time call by the decision-making system.

[0028] Specifically, the specific process of generating a dynamic constraint boundary table with variable priorities is as follows: based on the dynamic competitive characteristic matrix, the constraint intervals of flood control, power generation, and ecology are defined respectively, including the lower limit of flood control reservoir capacity demand, the upper limit of the feasible range of power generation output, and the rigid threshold of ecological flow; when the rainfall forecast intensity exceeds the historical threshold for the same period, the priority of the lower limit of flood control reservoir capacity is increased and the upper limit interval of power generation output is compressed; when the ecological sensor detects abnormal dissolved oxygen or fish activity, the ecological flow threshold is set as an insurmountable rigid constraint and the upper limit of power generation output is simultaneously lowered; if there is no overlap between the flood control and power generation constraint intervals, the upper limit of power generation output is recalculated with flood control as the highest priority; the adjusted constraint intervals and priority labels are encoded into a dynamic constraint boundary table.

[0029] In this implementation plan, first, based on the dynamic competitive feature matrix, the constraint intervals for flood control, power generation, and ecological objectives are defined respectively. The flood control constraint interval calculates the lower limit of the flood control storage capacity requirement through meteorological forecast data and hydrological monitoring data. For example, by combining future rainfall predictions and the current water level, a certain storage capacity is reserved to cope with flood risks. The power generation constraint interval derives the upper limit of the feasible power generation output according to the grid load curve and the unit efficiency model to ensure that while meeting the electricity demand, the significant decline in unit efficiency is avoided. The ecological constraint interval dynamically generates the minimum ecological flow threshold through real-time ecological sensor data (such as dissolved oxygen concentration, fish migration activities) and ecological response models. For example, during the spawning period, a specific flow rate needs to be maintained to protect fish habitats. When the predicted rainfall intensity exceeds the threshold of the same historical period, the system raises the priority of the flood control storage capacity requirement, and forces the reservation of more storage capacity by compressing the upper limit of the feasible power generation interval. If the dissolved oxygen concentration monitored by the ecological sensor is lower than the threshold or the fish activities are abnormal (such as spawning being blocked), the ecological flow threshold will be set as a non-breachable rigid constraint, and at the same time, the upper limit of the power generation output will be reduced synchronously to ensure that the ecological flow meets the standard. In addition, if there is no overlap in the storage capacity requirement intervals for flood control and power generation (for example, the occupancy rate of the flood control storage capacity is too high, resulting in insufficient available storage capacity for power generation), the system takes flood control as the highest priority, recalculates the maximum feasible value of the power generation output, and gives priority to ensuring flood control safety. The adjusted constraint intervals (the lower limit of the flood control storage capacity, the upper limit of the power generation output, the ecological flow threshold) and the dynamic priority labels (such as "rigid", "high", "medium", "low") are encoded into a dynamic constraint boundary table. This table provides a variable constraint boundary basis for multi-objective collaborative optimization by quantifying the conflict intensity (such as the storage capacity overlap rate) and the real-time priority status of each objective, and supports the dispatching decision-making system to flexibly balance the flood control, power generation, and ecological objectives in different scenarios.

[0030] Specifically, according to the dynamic constraint boundary table, the specific process of generating a Pareto feasible solution set that meets flood control safety, power generation benefits, and ecological flow compliance through the physical constraint double-depth network algorithm and matching multi-objective collaborative rules is as follows: Define the action space based on the lower limit of the flood control storage capacity requirement, the upper limit of the feasible power generation interval, and the rigid ecological flow threshold in the dynamic constraint boundary table; Embed the hydrodynamic equation as a physical constraint into the network state transition rule to limit the flood discharge flow rate not to exceed the safe capacity of the downstream river channel; Design a multi-objective reward function, including a penalty term for flood control water level control error, a gain term for power generation benefits, and a reward term for the ecological flow compliance rate; Optimize the network parameters through offline historical data pre-training and online real-time data fine-tuning, and output a Pareto feasible solution set that meets the dynamic constraints of flood control safety, power generation benefits, and ecological flow.

[0031] In this implementation plan, based on the flood control, power generation, and ecological constraint conditions in the dynamic constraint boundary table, the range of decision variables is set: the flood discharge flow rate ( ): ; : The maximum safe flood discharge of the downstream river channel is determined by the reservoir's carrying capacity. The power generation output ( ): ; : The upper limit of power generation output defined in the dynamic constraint boundary table is determined jointly by the grid load demand and the unit efficiency curve. The ecological discharge flow ( ): ; : The rigid threshold of ecological flow, such as the minimum flow for dissolved oxygen or fish migration requirements. Embed the hydrodynamic equation as a physical rule into the network state transition process to ensure the physical feasibility of flood discharge and power generation operations: Water balance equation: ; : The real-time inflow into the reservoir, obtained through hydrological monitoring data; : The power generation diversion flow, related to the power generation output through the water energy conversion coefficient κ ( ); V: The real-time reservoir storage capacity, which needs to be updated dynamically. Storage capacity constraint: ; : The lower limit of flood control storage capacity demand, dynamically calculated from rainfall prediction and hydrological models. Design a multi-objective reward function to quantify the achievement degrees of flood control, power generation, and ecological objectives through independent reward terms and adjust the priorities: Flood control water level control penalty term: ; : The upper limit of the storage capacity corresponding to downstream flood control safety, exceeding which will trigger flood control risks; : The flood control penalty weight coefficient, controlling the penalty intensity for exceeding the storage capacity limit. Power generation revenue gain term: ; : The time length of the scheduling period; : The unit power revenue price; : The power generation revenue weight coefficient, adjusting the importance of the power generation objective. Ecological flow compliance reward term: ; : The actually discharged ecological flow, monitored in real time through sensors; : The ecological objective weight coefficient, reflecting the priority of ecological flow compliance. Offline pre-training and online fine-tuning, offline prediction training: Use the historical data set ( ) to train the network, balancing the loss function ( ) and the multi-objective reward. Online fine-tuning: Combine real-time rainfall prediction ( ). Output the Pareto feasible solution set to generate a non-dominated solution set that satisfies the following constraints: ; Screening rule: Eliminate the invalid solutions that are comprehensively surpassed by other solutions, and retain the solutions that cannot be further optimized in any of the flood control, power generation, and ecological objectives.

[0032] Specifically, the specific process of mapping the solution set to the collaborative strategy of flood discharge flow and power generation output is as follows: Analyze the flood discharge flow, power generation output, and ecological water release instructions in the Pareto feasible solution set, and extract the maximum flood discharge flow, peak power generation output during the period, and the period when the ecological flow reaches the standard; Based on the current reservoir gate state and unit operating conditions, use the pre-trained hydraulic look-up table model to map the flood discharge flow to the gate opening control instructions including flow rate, corresponding opening, and response time; Combine the real-time load demand of the power grid and the unit ramp rate limit to generate a power generation output curve for each period; Monitor the downstream flow in real time according to the rigid threshold of the ecological flow. If the actual value is lower than the threshold, dynamically increase the opening of the water release gate according to the shortage ratio and synchronously reduce the upper limit of power generation output; Encode the flood discharge gate opening, power generation output curve, and ecological water release instructions into a collaborative control instruction sequence according to the time stamp, and push it to the reservoir execution terminal through the industrial communication protocol.

[0033] In this implementation plan, based on the pre-trained hydraulic look-up table model, the flood discharge flow is converted into gate control instructions: Parameter description: : The flood discharge flow in the k-th period; : The current gate state (such as opening, response speed); : The hydraulic look-up table model, which inputs the flow rate and gate state and outputs the target opening; : The target gate opening in the k-th period. Control logic: According to the flow rate-opening relationship table (for example: 100 m³ / s corresponds to an opening of 50%), generate a stepped opening instruction and mark the response time Δt (such as it takes 10 minutes to adjust the opening). Generate the power generation output curve for each period, and combine the power grid demand and the unit ramp rate limit to generate a power generation plan: Parameter description: : The power grid load demand in the t-th period; : The upper limit of unit output; : The unit ramp rate, which limits the output change speed; : The length of the scheduling period (unit: min); : The power generation output in the previous period. Logic: On the premise of meeting the ramp rate limit, give priority to responding to peak loads (such as running at full load). Dynamically adjust the ecological flow. Monitor the downstream ecological flow in real time and trigger rigid constraint adjustment: (if ) Parameter description: : The actually monitored ecological flow; : The flow shortage; : The gate opening compensation coefficient (for example: for every 1 m³ / s shortage, the opening increases by 2%); : Power generation output reduction coefficient (e.g., for every 1 m^3 / s shortage, the output decreases by 0.5 MW). Logic: If the ecological flow is insufficient, the gate opening is preferentially increased and the power generation output is decreased to ensure . The coding collaborative control instruction sequence integrates each parameter into an executable instruction according to the timestamp: instruction Parameter description: : The timestamp of the th time period; : The adjusted gate opening; : The corrected power generation output; : Ecological water discharge instruction. Transmission protocol: Push it to the execution terminal through an industrial communication protocol (such as ModbusTCP), and the instruction interval is ΔT (such as 15 minutes).

[0034] Specifically, based on the Pareto feasible solution set, the specific process of dynamically adjusting the flood control and power generation weights through the visualization interface is as follows: Load the lower limit of the flood control storage capacity demand, the upper limit of the power generation benefit, and the rigid threshold of the ecological flow in the dynamic constraint boundary table into the visualization interface; Adjust the flood control weight and the power generation weight through drag-and-drop slider interaction to generate the weight distribution ratio; Recalculate the Pareto front based on the adjusted weights and screen the subset of the solution set that meets the weight interval; Generate a real-time heat map of the flood control risk level.

[0035] In this implementation plan, the system first loads a dynamic constraint boundary table including key parameters such as the lower limit of flood control storage capacity demand, the upper limit of power generation benefit, and the rigid threshold of ecological flow. By dragging the slider to interactively adjust the flood control weight and power generation weight. In this part, the user performs interactive operations through the slider in the visual interface to dynamically adjust the weights of flood control and power generation goals. This method enables the user to flexibly adjust the importance of the two goals according to actual needs, ensuring the flexibility of the system. Flood control weight: It represents the relative importance of the flood control goal in the overall reservoir operation. The user adjusts this weight by sliding the slider, determining the priority of the flood control goal in the final operation decision. Power generation weight: Corresponding to the flood control weight, the power generation weight is also dynamically adjusted through the slider, representing the importance of the power generation goal. The sum of the two is usually 1, that is, the sum of the flood control weight and the power generation weight does not exceed 100%. By adjusting these two weights, the user can balance the relationship between flood control and power generation, ensuring that the operation decision meets the actual needs. After the user adjusts the flood control and power generation weights, the system will recalculate the Pareto front according to the new weights. The Pareto front represents all the optimal solution sets between flood control and power generation, that is, maximizing one goal without sacrificing the other. Pareto front calculation: According to the new weights, the system will recombine the existing constraints (such as flood control storage capacity demand, upper limit of power generation benefit, ecological flow threshold, etc.) to recalculate all possible operation plans and screen out the optimal solution set under the current weights. In this process, the interaction between flood control and power generation will be accurately calculated, and a new optimal solution set will be obtained. Recalculate the Pareto front and screen the subset of solution sets. Based on the adjusted weights, screen the subset of the Pareto solution set that meets the weight interval: Weighted sum of objective functions: ; Flood control safety reward (the closer the storage capacity is to the safety value, the higher the reward); : Power generation revenue reward (the closer the output is to the upper limit, the higher the reward); : Ecological compliance reward (the closer the flow is to the threshold, the higher the reward). Screening rule: Retain The top 10% of the solutions are selected, and those with poor performance in low-weight objectives are excluded. After calculating the Pareto front, the system will further screen out the subset of the solution set that meets the current adjusted weight range. That is, among all the calculated optimal solutions, those that still meet the constraint conditions under the adjusted flood control and power generation weight ratios are selected. Screening process: The system will screen out those solutions that balance the multi-objective requirements such as flood control safety, power generation benefits, and ecological flow according to the flood control and power generation weights set by the user. This screening process ensures that each selected solution set can meet the specific needs of the user. Finally, based on the screened subset of the Pareto solution set, the system will generate a real-time heat map of the flood control risk level. The heat map provides intuitive visual support for decision-makers by calculating the flood control risks in different regions. The flood control risk levels in different regions will be dynamically generated according to the dispatching scheme after weight adjustment and presented in different colors or color scales.

[0036] Specifically, in combination with the hydraulic model to pre-deduce the impact path of flood discharge on the downstream inundated area and ecological flow of the reservoir, the specific process of generating executable dispatching instructions is as follows: Drive the hydraulic model according to the hydrological monitoring data, the rigid threshold of ecological flow, and the collaborative control instruction sequence, simulate the propagation path of flood discharge flow and the inundation range, predict the maximum water depth and duration of the downstream inundated area of the reservoir, and evaluate whether the dissolved oxygen recovery time and the compliance rate of fish habitats meet the dynamic constraint boundary table; if not, regenerate the Pareto solution set and update the collaborative control instruction sequence; if satisfied, push the verified collaborative control instruction sequence encoded by the time stamp to the execution terminal.

[0037] In this implementation plan, the hydraulic model is driven by hydrological monitoring data, the rigid threshold of ecological flow, and the collaborative control instruction sequence. At this stage, the system first drives the hydraulic model for simulation based on real-time hydrological monitoring data, the rigid threshold of ecological flow, and the collaborative control instruction sequence generated in the early stage. Hydrological monitoring data usually includes information such as precipitation, flow rate, and water level; the rigid threshold of ecological flow is used to ensure the minimum water flow required for the downstream ecosystem to be satisfied; and the collaborative control instruction sequence is a collaborative strategy for flood discharge and power generation generated based on the multi-objective optimization algorithm. The role of the hydraulic model: The hydraulic model is used here to simulate the propagation path and inundation range of the reservoir flood discharge flow rate, and predict the impact of different flood discharge strategies on the downstream area, including the rise of water level and the expansion of inundation. Driven by the hydraulic model, the system simulates the propagation path of the water flow downstream under different flood discharge flow rate conditions. Through numerical simulation and fluid mechanics calculation, the hydraulic model predicts the expansion range of the flood discharge flow rate and determines the specific range and change trend of the inundated area. Prediction of inundated area: The system determines which areas downstream may be inundated after flood discharge by calculating the relationship between the flood discharge flow rate and factors such as terrain and water level, and evaluates the flow direction and water accumulation depth of the water flow under different flood discharge volumes. To predict the maximum water depth and duration of the inundated area downstream of the reservoir, the hydraulic model will further predict the maximum water depth and duration of the inundated area downstream under different flood discharge scenarios. These parameters are crucial for flood control safety and can help decision-makers evaluate the potential impact of flood discharge and formulate corresponding countermeasures. Maximum water depth: By simulating the water flow path and the accumulation effect of the water flow, the system calculates the maximum water depth that may occur after flood discharge. This helps to evaluate which low-lying areas or important facilities may be threatened. Duration: Predict the duration of the water level, usually calculated based on factors such as the flood discharge volume, the drainage capacity of the basin, and the downstream terrain. The estimation of the duration helps decision-makers determine whether the impact of the inundated area is short-term or long-term. Evaluate the dissolved oxygen recovery time and the compliance rate of fish habitats. While ensuring flood control safety, it is also necessary to evaluate the impact of ecological flow on the ecosystem. The hydraulic model simulates the change in dissolved oxygen level and the recovery of fish habitats in the downstream area according to the flow rate change. Dissolved oxygen recovery time: According to the flow rate change and water flow dynamics, the system calculates the recovery time of the dissolved oxygen concentration after flood discharge. If the flood discharge volume is too large, the dissolved oxygen may drop suddenly, which will affect the water quality and ecological environment. Compliance rate of fish habitats: Simulate whether the fish habitats will be damaged and predict the recovery of habitats under different flood discharge scenarios. Through these evaluations, ensure that the rigid threshold of ecological flow is not breached. Compare the above prediction results with the dynamic constraint boundary table. The dynamic constraint boundary table contains the priorities and acceptable thresholds among flood control, power generation, and ecological flow.If not satisfied: If the simulation results show that the flood discharge plan does not meet the requirements of multi-objectives such as flood control, power generation, and ecological flow (such as too long dissolved oxygen recovery time, inability to restore fish habitats), the system needs to regenerate the Pareto solution set and update the collaborative control instruction sequence to meet the new constraint conditions and objective requirements. If satisfied: If the simulation results indicate that the flood discharge plan meets all objective constraints (flood control safety, power generation benefits, ecological flow, etc.), the final executable scheduling instructions can be generated. Once verified, the system generates the final scheduling instructions and encodes them with timestamps to ensure their execution in the correct chronological order. These scheduling instructions include control information such as the adjustment of flood discharge flow and the operating status of generator sets. Timestamp encoding: Each instruction is accompanied by a clear timestamp to ensure that the instructions can be executed in the set chronological order and priority to avoid conflicts. Pushed to the execution terminal: The scheduling instructions are pushed to specific execution terminals (such as reservoir scheduling control systems) through the network, and these terminals will perform specific scheduling operations according to the instructions.

[0038] Specifically, the specific process of correcting the dynamic constraint boundary table according to the actual ecological flow monitoring data is as follows: The actual ecological flow data downstream of the reservoir is collected in real time through ecological sensors. If the monitoring values for a continuously set duration are lower than the ecological flow rigid threshold defined in the dynamic constraint boundary table, it is determined that the ecological constraint is not met, the priority of the ecological flow constraint is increased, and the lower limit of the flood control storage capacity requirement and the upper limit of the feasible interval of power generation output are compressed, triggering the regeneration of the Pareto feasible solution set and the update of the collaborative control instruction sequence.

[0039] In this implementation plan, the system collects ecological flow data downstream of the reservoir in real time through ecological sensors. Usually, these sensors monitor ecological data such as water flow velocity, dissolved oxygen concentration, and fish migration. These data reflect the health status of the ecological system downstream of the reservoir. According to the set time period, if the monitored data is lower than the ecological flow rigid threshold defined in the dynamic constraint boundary table for a continuous period of time, it is considered that the ecological flow does not meet the predetermined standard. This judgment is based on the minimum value of the flow data, that is, if the minimum flow is lower than the threshold within the set time range, it is considered that the ecological flow does not meet the standard. If the ecological flow does not meet the standard, the system will adjust the strategy according to the priority, and raise the ecological flow constraint to a higher priority. This means that the system will give priority to ensuring that the ecological flow demand is met, and may thus reduce the priorities of other objectives (such as flood control or power generation). For example, the lower limit of the flood control storage capacity requirement may be compressed, and the upper limit of the power generation output may be reduced, so as to release more water to meet the ecological flow demand. Due to the increase in the priority of the ecological flow, the constraint conditions for flood control and power generation have changed, and the system needs to recalculate the new Pareto feasible solution set. The Pareto solution set is the solution set that finds the optimal balance point among multiple objectives (such as flood control, power generation, and ecological protection). When regenerating the solution set, the system will calculate the optimal reservoir operation strategy according to the new constraint conditions (such as the adjusted flood control storage capacity requirement and the upper limit of power generation). When the priority of the ecological flow constraint is increased and the flood control and power generation intervals are compressed, the system needs to regenerate the Pareto feasible solution set and update the collaborative control instruction sequence. This is because the solution set of the optimization problem will change due to the change of the constraint conditions. The formula: ; : The corrected Pareto feasible solution set P: The Pareto optimization solution set function; : The corrected lower limit of the flood control storage capacity requirement; : The corrected upper limit of the feasible interval of the power generation output; : The target ecological flow. When the new Pareto solution set is generated, the system will update the collaborative control instruction sequence. This instruction sequence will be encoded with timestamps to ensure that the instructions are executed in the correct chronological order. Finally, these updated scheduling instructions will be pushed to the execution terminal to ensure that the actual operation is scheduled according to the latest instructions, so as to realize the optimization of the reservoir operation.

[0040] In summary, this application has at least the following effects:

[0041] A reservoir operation decision support system for multi-objective optimization can effectively coordinate the conflicts among multiple objectives by integrating the dynamic constraints and priority adjustments of three major objectives: flood control, power generation, and ecological flow. It ensures the optimal balance of each objective under different circumstances and enhances the comprehensive benefits of reservoir operation. By increasing the priority of ecological flow constraints and real-time correcting the dynamic constraint boundary table, the system can ensure that the ecological flow requirements are preferentially met, effectively protect the ecological environment downstream of the reservoir, and avoid ecological damage. Using blockchain technology to record the evidence of the dispatching decision-making process ensures the transparency and traceability of the dispatching process, and enhances the credibility and security of the decision-making. Through the preview of the flood discharge path and inundated area based on the hydraulic model, the system can predict in advance the impact of flood discharge on the downstream area, thereby optimizing the reservoir operation plan and reducing unnecessary resource waste and potential risks. Using the Pareto front optimization and cooperative control algorithms, the system can automatically generate the best dispatching instructions according to the priorities of each objective and push them for execution in real time to ensure the efficiency and security of reservoir operation.

[0042] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0043] The present invention is described with reference to the flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0044] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0045] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one process or a plurality of processes and / or blocks Figure 1 one process or a plurality of processes and / or blocks Figure 1 or steps of the functions specified in a plurality of blocks.

[0046] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0047] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A reservoir operation decision support system for multi-objective optimization, characterized in that Includes the following modules: Multi-source dynamic constraint analysis module, collaborative optimization module, scheduling module, and correction module; The multi-source dynamic constraint analysis module is used to access meteorological forecast data, hydrological monitoring data and ecological sensor data in real time through the Internet of Things, extract the competitive characteristics of flood control, power generation and ecological goals through dynamic threshold calculation, and generate a dynamic constraint boundary table with variable priority; The collaborative optimization module is used to match multi-objective collaborative rules according to the dynamic constraint boundary table through the physical constraint double-depth network algorithm to generate a Pareto feasible solution set that meets the flood control safety, power generation income and ecological flow standards, and map the solution set into a collaborative strategy for flood discharge and power generation output; The scheduling module is used to dynamically adjust the weights of flood control and power generation through a visual interface based on the Pareto feasible solution set, and to generate executable scheduling instructions by combining a hydraulic model to simulate the impact path of flood discharge on the flooded area and ecological flow downstream of the reservoir; The correction module is used to correct the dynamic constraint boundary table according to the actual ecological flow monitoring data based on the executable scheduling instructions and the scheduling decision process recorded through the blockchain.

2. The reservoir operation decision support system for multi-objective optimization according to claim 1, characterized in that: The specific process of extracting the competitive characteristics of flood control, power generation and ecological goals through dynamic threshold calculation is as follows: The lower limit of flood control reservoir capacity demand is calculated based on meteorological forecast data and hydrological monitoring data, and the upper limit of power generation efficiency is derived by combining the grid load curve and unit efficiency model. At the same time, the minimum ecological flow threshold is dynamically generated based on the dissolved oxygen and fish migration data monitored by ecological sensors; By comparing the overlap rate between flood control reservoir capacity demand and power generation benefit range, the conflict intensity between flood control and power generation goals is determined. When the ecological sensor detects that the dissolved oxygen is lower than the threshold or the fish activity is abnormal, the priority of ecological flow constraints is increased, and the calculation results are output as a dynamic competitive characteristic matrix of flood control, power generation, and ecological goals.

3. The reservoir operation decision support system for multi-objective optimization according to claim 2, wherein: The specific process of generating a dynamic constraint boundary table with variable priority is as follows: Based on the dynamic competitive characteristic matrix, the constraint intervals of flood control, power generation, and ecology are defined respectively, including the lower limit of flood control reservoir capacity demand, the upper limit of power generation output feasible interval, and the rigid threshold of ecological flow; When the rainfall forecast intensity exceeds the historical threshold for the same period, the lower limit of flood control reservoir capacity is prioritized and the upper limit of power generation output is compressed; When the ecological sensor detects abnormal dissolved oxygen or fish activity, the ecological flow threshold is set as an insurmountable rigid constraint and the upper limit of power generation output is simultaneously lowered; If there is no overlap between the flood control and power generation constraint intervals, the power generation output limit is recalculated with flood control as the highest priority; The adjusted constraint intervals and priority labels are encoded into a dynamic constraint boundary table.

4. The reservoir operation decision support system for multi-objective optimization according to claim 3, characterized in that: According to the dynamic constraint boundary table, the specific process of matching multi-objective coordination rules through the physical constraint dual-depth network algorithm to generate a Pareto feasible solution set that meets the flood control safety, power generation benefits and ecological flow standards is as follows: The action space is defined based on the lower limit of flood control reservoir capacity demand, the upper limit of feasible power generation output range and the rigid threshold of ecological flow in the dynamic constraint boundary table; The hydrodynamic equations are embedded as physical constraints in the network state transition rules to limit the flood discharge to no more than the safety capacity of the downstream river channel. Design a multi-objective reward function, including a penalty term for flood control water level control error, a gain term for power generation revenue, and a reward term for the ecological flow compliance rate; Optimize the network parameters through offline historical data pre-training and online real-time data fine-tuning, and output a Pareto feasible solution set that meets the dynamic constraints of flood control safety, power generation revenue, and ecological flow.

5. The reservoir operation decision support system for multi-objective optimization according to claim 4, characterized in that: The specific process of mapping the solution set to the coordinated strategy of flood discharge flow and power generation output is as follows: Analyze the flood discharge flow, power generation output, and ecological water release instructions in the Pareto feasible solution set, and extract the maximum flood discharge flow, peak power generation output during the period, and ecological flow compliance period; Based on the current reservoir gate state and unit operating conditions, map the flood discharge flow to the gate opening control instructions including flow, corresponding opening, and response time through a pre-trained hydraulic look-up table model; Combine the real-time load demand of the power grid and the unit ramp rate limit to generate a power generation output curve in sub-periods; Monitor the downstream flow in real time according to the rigid threshold of ecological flow. If the actual value is lower than the threshold, dynamically increase the opening of the water release gate according to the shortage ratio and synchronously reduce the upper limit of power generation output; Encode the flood discharge gate opening, power generation output curve, and ecological water release instructions into a coordinated control instruction sequence according to the time stamp, and push it to the reservoir execution terminal through the industrial communication protocol.

6. The reservoir operation decision support system for multi-objective optimization according to claim 5, characterized in that: Based on the Pareto feasible solution set, the specific process of dynamically adjusting the flood control and power generation weights through the visualization interface is as follows: Load the lower limit of flood control storage capacity demand, the upper limit of power generation benefit, and the rigid threshold of ecological flow in the dynamic constraint boundary table to the visualization interface; Adjust the flood control weight and power generation weight through drag-and-drop slider interaction to generate a weight allocation ratio; Recalculate the Pareto front based on the adjusted weights, and screen the subset of the solution set that meets the weight interval; Generate a heat map of flood control risk levels in real time.

7. The reservoir operation decision support system for multi-objective optimization according to claim 6, characterized in that: The specific process of generating executable scheduling instructions by combining the hydraulic model to pre-demonstrate the impact path of flood discharge on the downstream submerged area and ecological flow of the reservoir is as follows: Drive the hydraulic model according to the hydrological monitoring data, the rigid threshold of ecological flow, and the coordinated control instruction sequence, simulate the flood discharge flow propagation path and inundation range, predict the maximum water depth and duration of the downstream submerged area of the reservoir, and evaluate the dissolved oxygen recovery time and the compliance rate of fish habitats to see if they meet the dynamic constraint boundary table; If not satisfied, regenerate the Pareto solution set and update the coordinated control instruction sequence; If satisfied, push the verified coordinated control instruction sequence encoded according to the time stamp to the execution terminal.

8. The reservoir operation decision support system for multi-objective optimization according to claim 7, characterized in that: The specific process of correcting the dynamic constraint boundary table according to the actual ecological flow monitoring data is as follows: Real-time collect the actual ecological flow data downstream of the reservoir through ecological sensors. If the monitoring value is lower than the rigid threshold of ecological flow defined in the dynamic constraint boundary table for a continuous set duration, it is determined that the ecological constraint is not met, the priority of the ecological flow constraint is increased, and the lower limit of the flood control storage capacity demand and the upper limit of the feasible interval of power generation output are compressed, triggering the regeneration of the Pareto feasible solution set and the update of the coordinated control instruction sequence.

Images