Reservoir flood season sediment optimization dispatching method under cascade reservoir combined dispatching

Abstract

The invention provides a reservoir flood season sediment optimization scheduling method under cascade reservoir combined scheduling. The method comprises the following steps: constructing a dynamic model; adjusting scheduling starting conditions according to different water sand characteristics; screening out a scheme set with relatively high comprehensive benefits; starting traceability scouring and density flow desilting cooperative scheduling; adjusting a scheduling scheme in real time according to the monitoring condition; and comprehensively evaluating the scheduling effect. The traceability scouring is set, the water level in front of the dam rapidly drops, a large gradient and an obvious drop sill can be formed with the vertex of upstream silt in a short time, a large amount of silt at the bottom of the reservoir can be effectively taken up along with continuous development of the drop sill towards the upstream, the silt can be discharged out of the reservoir, the silt deposition amount in the reservoir is remarkably reduced, and the service life of the reservoir is prolonged. And the problems of reservoir capacity reduction, reservoir function decline and the like caused by sediment deposition are reduced.

Description

Technical Field

[0001] The present invention relates to the field of water conservancy projects, and particularly to an optimized sediment regulation method for reservoirs during the flood season under the joint operation of cascade reservoirs. Background Technique

[0002] In the construction and operation of water conservancy projects, the role of reservoirs is crucial. They can not only regulate the temporal and spatial distribution of water resources but also play a key role in flood control, power generation, irrigation, water supply, and ecological protection. Especially for cascade reservoir groups, their joint operation can achieve more efficient utilization of water resources. However, the sediment problem in reservoirs during the flood season has always been a key factor affecting the long-term stable operation of reservoirs and the exertion of comprehensive benefits.

[0003] During the flood season, a large amount of sediment is carried into the reservoir by the river. The sediment deposition in the reservoir will lead to a reduction in the reservoir capacity, a decrease in the regulation capacity of the reservoir, and an impact on functions such as flood control and power generation. At the same time, unreasonable sediment regulation may have a negative impact on the downstream ecological environment, shipping conditions, etc. For example, excessive sediment discharge may cause scouring of the downstream river channel, affecting the stability of riverbanks and ecological balance; while too much sediment deposition will reduce the navigability of the river channel.

[0004] At present, certain achievements have been made in the existing technology for sediment regulation in reservoirs during the flood season. The existing methods mainly rely on historical experience and simple models for scheduling decisions. For example, some methods set fixed scheduling thresholds based on multi-year average water and sediment data, and when the inflow or sediment concentration reaches the threshold, the corresponding scheduling plan is initiated. There are also some methods that use single-objective optimization models, only considering one aspect of the goal such as flood control or sediment reduction, ignoring the mutual influence and collaborative requirements among multiple objectives in reservoir operation. Summary of the Invention

[0005] The present application provides an optimized sediment regulation method for reservoirs during the flood season under the joint operation of cascade reservoirs to solve the problems raised in the above background technique. The object of the present invention is achieved as follows: An optimized sediment regulation method for reservoirs during the flood season under the joint operation of cascade reservoirs includes the following steps: Step S1: Collect multi-source data related to sediment regulation in reservoirs during the flood season, perform fusion processing and analysis on the collected multi-source data, and construct a dynamic model reflecting the current water and sediment conditions and the interaction relationship between factors; Step S2: Based on the dynamic model in Step S1 and historical multi-source data, adopt an adaptive optimization method to adjust the scheduling start conditions according to different water and sediment characteristics; Step S3: Before making a scheduling decision, pre-simulate different scheduling plans, conduct a risk assessment on the pre-simulated plans based on the mutual influence and complex conditions of the reservoir group, and screen out a set of plans whose comprehensive benefits meet the requirements, that is, screen out a set of plans with higher comprehensive benefits; Step S4: According to the real-time monitoring and pre-simulation results, initiate the collaborative scheduling of headward erosion and density current sediment discharge, and monitor and adjust the relevant parameters of headward erosion and density current sediment discharge during the scheduling process; Step S5: During the collaborative sediment discharge scheduling process, monitor the operation indicators of the reservoir group and multiple downstream targets in real time, achieve the dynamic balance of each target based on multi-objective optimization, and adjust the scheduling plan in real time according to the monitoring situation; Step S6: After the scheduling is completed, comprehensively evaluate the scheduling effect, and feedback the evaluation results to the above-mentioned dynamic model and scheduling start condition setting link for learning and optimization.

[0006] In the present invention, in the step S1, the multi-source data includes reservoir hydrological data, meteorological data, basin land use change data, and reservoir surrounding topographic and geomorphic data; the multi-source data is fused using a weighted average fusion function, and the formula is as follows: ; where, is the fused data value, is the data value of the th data source, is the weight corresponding to the th data source, and , is the number of data source types.

[0007] In the present invention, the relevant parameters of headward erosion and density current sediment discharge include the flow rate and sediment concentration data of the main control stations for inflow and outflow of multiple reservoirs in the upstream and downstream of a river or river section and other hydrological stations in the reservoir area.

[0008] In the present invention, in the step S2, the adaptive optimization method adopts a dynamic programming algorithm combined with a machine learning algorithm. The dynamic programming is used to adjust the scheduling start conditions according to different water and sediment characteristics. The one-dimensional dynamic programming recurrence formula is as follows: ; where, is the optimal value function when the state is in stage , is the decision variable in stage , is the set of allowable decisions in state , is the immediate benefit obtained by taking decision from state in stage , is the discount factor, is from state and decision The determined next-stage state.

[0009] In the present invention, in the step S3, the complex situations include extreme water and sediment conditions. A risk assessment is carried out on the pre-simulation scheme, and a linear weighted combination method is used to evaluate multiple risk indicators. The formula is as follows: ; Wherein, is the comprehensive risk value, is the th risk indicator value, is the weight of the th risk indicator, is the number of risk indicators.

[0010] In the present invention, in the step S4, the initiation of headward erosion is achieved by controlling the upstream reservoir to adjust the water level in front of the dam, forming a drop-off with a height difference from the upstream silt apex, and monitoring the movement of sediment at the bottom of the reservoir and the change of the water level in front of the dam. The density current sediment discharge is achieved by adjusting the reservoir discharge mode, inducing the muddy water to form a stable density current moving towards the dam at the bottom of the reservoir, and monitoring the relevant parameters of the density current.

[0011] In the present invention, the adjustment of the reservoir discharge mode includes adjusting the discharge flow rate and the position of the discharge outlet.

[0012] In the present invention, the relevant parameters of the headward erosion and the density current sediment discharge include the movement speed of sediment at the bottom of the reservoir, the change rate of the water level in front of the dam, the shape, speed, and sediment concentration of the density current.

[0013] In the present invention, in the step S5, the multi-objective optimization is realized based on a multi-objective optimization algorithm. The multi-objectives include flood control, sediment reduction, ecology, and water resource utilization. The weighted sum method is used based on the multi-objective optimization algorithm. The formula is as follows: ; Wherein, is the comprehensive objective function value, is the th objective function, is the weight of the th objective function, is the decision variable vector, is the number of objective functions.

[0014] In the present invention, in the step S6, the indicators for comprehensive evaluation include the change in sediment deposition volume, the improvement of flood control ability, the improvement of ecological environment, the change in shipping conditions, and the water resource utilization efficiency. The learning and optimization use a machine learning algorithm to process the feedback data. The machine learning algorithm uses a linear regression model. The formula is as follows: ; Among them, is the predicted value, is the intercept, is the regression coefficient, is the independent variable.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: By setting up the headward erosion scouring, the water level in front of the dam drops rapidly, and a large slope and an obvious drop can be formed with the apex of the upstream silt in a short time. As the drop continuously develops upstream, a large amount of silt deposited at the reservoir bottom can be effectively lifted, discharged out of the reservoir, significantly reducing the silt deposition amount in the reservoir, prolonging the service life of the reservoir, and reducing problems such as the reduction of reservoir capacity and the decline of reservoir functions caused by silt deposition. The density current sediment discharge utilizes the characteristics of the stratified flow of turbid water and clear water. The high-sediment-concentration water flow close to the reservoir bottom drives the sediment at the reservoir bottom to move together during the movement process and evolves to be discharged out of the reservoir in front of the dam, further enhancing the sediment discharge capacity of the reservoir. Acting together from different mechanisms, efficient silt reduction is achieved; The coordinated operation of the headward erosion scouring and the density current sediment discharge of the present invention can effectively reduce the silt deposition in the reservoir in a timely manner, reduce the flood risk increased by the rising water level due to silt deposition during the flood season of the reservoir. Through reasonable scheduling, the reservoir is maintained with sufficient regulating storage capacity to ensure that the reservoir can give full play to the role of storing floodwaters when floods come, and guarantee the flood control safety of the upstream and downstream areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will be further described below in conjunction with the drawings and embodiments.

[0017] Figure 1 is a schematic flow chart of the method for optimizing sediment regulation in the flood season of a reservoir under the joint operation of cascade reservoirs of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] Please refer to Figure 1 , the present invention provides a technical solution: A method for optimizing sediment regulation in the flood season of a reservoir under the joint operation of cascade reservoirs, comprising the following steps: Step S1: Collect multi-source data related to sediment regulation in the flood season of the reservoir, perform fusion processing and analysis on the collected multi-source data, and construct a dynamic model reflecting the current water and sediment conditions and the interaction relationship between factors; In step S1, the multi-source data includes reservoir hydrological data, meteorological data, basin land use change data, and reservoir surrounding topographic and geomorphic data; the multi-source data is fused using a weighted average fusion function, and the formula is as follows: ; Among them, is the fused data value, is the data value of the th data source, is the weight corresponding to the th data source, and , is the number of data source types.

[0020] The reservoir hydrological data includes the flow and sediment concentration data of the main control stations for inflow and outflow of multiple reservoirs upstream and downstream of a river or river section, as well as other hydrological stations in the reservoir area.

[0021] In a specific embodiment, in a cascade reservoir system including three reservoirs, it is necessary to fuse the flow, sediment concentration, and water level data from different monitoring points to more accurately understand the water and sediment conditions of the reservoirs and provide a basis for subsequent scheduling.

[0022] Suppose the currently collected data is: The flow data of reservoir A , the sediment concentration data , and the water level data .

[0023] The flow data of reservoir B , the sediment concentration data , and the water level data .

[0024] The flow data of reservoir C , the sediment concentration data , and the water level data .

[0025] And based on past experience and expert judgment, the flow data weight , the sediment concentration data weight , and the water level data weight are determined.

[0026] Specifically: Step S101, first, clarify that the goal is to fuse the data from multiple data sources into a comprehensive data to more comprehensively reflect the water and sediment conditions of the reservoir system. Weighted average fusion is a simple and effective method that takes into account the importance (weight) of different data sources.

[0027] Step S102: For each data type (flow rate, sediment concentration, water level), calculate its weighted average respectively. Taking the flow rate as an example, there are a total of three reservoirs' flow rate data, and its weighted average is calculated as follows: The weighted value of the flow rate of Reservoir A is .

[0028] The weighted value of the flow rate of Reservoir B is .

[0029] The weighted value of the flow rate of Reservoir C is .

[0030] The fusion value of the flow rate data is .

[0031] Step S103: Similarly, for the sediment concentration data: The weighted value of the sediment concentration of Reservoir A is .

[0032] The weighted value of the sediment concentration of Reservoir B is .

[0033] The weighted value of the sediment concentration of Reservoir C is .

[0034] The fusion value of the sediment concentration data is .

[0035] Step S104: For the water level data: The weighted value of the water level of Reservoir A is .

[0036] The weighted value of the water level of Reservoir B is .

[0037] The weighted value of the water level of Reservoir C is .

[0038] The fusion value of the water level data is .

[0039] Step S105: The final fusion data is represented as a vector containing the fusion values of the flow rate, sediment concentration, and water level: .

[0040] Step S2, based on the dynamic model in Step S1 and historical multi-source data, adopts an adaptive optimization method to adjust the scheduling start condition according to different water and sediment characteristics; In Step S2, the adaptive optimization method adopts a dynamic programming algorithm combined with a machine learning algorithm. The dynamic programming is used to adjust the scheduling start condition according to different water and sediment characteristics. The one-dimensional dynamic programming recurrence formula is as follows: ; Among them, is the optimal value function at stage with the state of ; is the decision variable at stage ; is the set of allowable decisions at state ; is the immediate benefit obtained by taking decision from state at stage ; is the discount factor, is the next stage state determined by state and decision .

[0041] In a specific embodiment, assume that the above cascade reservoir system is scheduled within one month (with each day as a stage, a total of 30 stages, ). The inflow and sediment concentration of each stage are different. The reservoir needs to minimize sediment deposition as much as possible while ensuring flood control safety and meeting certain power generation requirements. The state variable of the reservoir includes the current water level and water storage capacity, and the decision variable is the daily discharge. Assume that on the th day, the initial water level of the reservoir is , the water storage capacity is , the inflow is , and the sediment concentration is . The immediate benefit function is defined as the power generation benefit minus the flood control risk loss and sediment deposition loss. The power generation benefit is related to the discharge and water level, the flood control risk loss is related to the probability that the reservoir water level exceeds the warning level, and the sediment deposition loss is related to the inflow sediment concentration and discharge. The discount factor .

[0042] Specifically: Step S201: The basic idea of dynamic programming is to decompose a complex multi-stage decision-making problem into a series of interrelated sub-problems and obtain the global optimal solution by solving the sub-problems.

[0043] Step S202: At the th stage, it is necessary to determine the optimal decision at state such that the total benefit from the th stage to the last stage is maximized.

[0044] Step S203: First, consider the immediate benefit , the power generation revenue is expressed as , where is the density of water, is the acceleration due to gravity, is the head corresponding to the water level, is the power generation efficiency. The flood control risk loss can be calculated by establishing a probability model of the water level exceeding the warning water level. Assuming the warning water level is , when , there is a flood control risk, and the risk loss is .

[0045] The sediment siltation loss is expressed as: (assuming that the siltation is proportional to the amount of sediment not discharged).

[0046] Then the immediate revenue: - .

[0047] Step S204: The state of the next stage can be obtained through the water balance equation: , .

[0048] Step S205: According to the recurrence formula of dynamic programming , start the derivation from the last stage . In the 30th stage, , because there is no subsequent stage, the discount factor does not need to be considered. Then calculate the optimal value function and the optimal decision of the 29th stage, the 28th stage... until the 1st stage in turn.

[0049] Step S3: Before making the scheduling decision, pre-simulate different scheduling schemes. Based on the mutual influence and complex conditions of the reservoir group, conduct a risk assessment on the pre-simulated schemes, and screen out the set of schemes that meet the requirements of the comprehensive benefit, that is, screen out the set of schemes with higher comprehensive benefits; In step S3, the complex conditions include extreme water and sediment conditions. When conducting a risk assessment on the pre-simulated schemes, a linear weighted combination method is used to evaluate multiple risk indicators, and the formula is as follows: ; where is the comprehensive risk value, is the value of the th risk indicator, is the weight of the th risk indicator, is the number of risk indicators.

[0050] In a specific embodiment, when pre-simulating and evaluating the above cascade reservoir operation plan, three main risk indicators are considered: flood control risk, ecological risk, and shipping risk. Assume that through historical data and model analysis, under the current operation plan, the flood control risk indicator value (indicating that there is a 30% probability that the reservoir water level will exceed the warning level), the ecological risk indicator value (indicating that the guarantee degree of downstream ecological flow is 80%, and the risk value is 1 - 0.8), and the shipping risk indicator value (indicating that there is a 10% probability of affecting downstream shipping). And according to the requirements of the reservoir management department, the flood control risk weight , the ecological risk weight , and the shipping risk weight .

[0051] Specifically: Step S301: The purpose of comprehensive risk assessment is to integrate multiple risk indicators of different types into a comprehensive risk value for evaluating the overall risk of the operation plan.

[0052] Step S302: The method of linear weighted combination is adopted because different risk indicators have different degrees of influence on the overall risk, and this difference is reflected by weights.

[0053] Step S303: According to the formula , multiply each risk indicator by its corresponding weight and then sum them up.

[0054] Step S304: The weighted value of flood control risk is .

[0055] Step S305: The weighted value of ecological risk is .

[0056] Step S306: The weighted value of shipping risk is .

[0057] Step S307: The comprehensive risk value .

[0058] Step S4: According to the real-time monitoring and pre-simulation results, start the collaborative scheduling of headward erosion and density current sediment discharge, and monitor and adjust the relevant parameters of headward erosion and density current sediment discharge during the scheduling process; In Step S4, the initiation of headward erosion is achieved by controlling the upstream reservoir to adjust the water level in front of the dam to form a drop with a difference from the upstream silt apex, and monitoring the movement of sediment at the bottom of the reservoir and the change of the water level in front of the dam. The density current sediment discharge is achieved by adjusting the reservoir discharge mode to induce the formation of a stable density current of muddy water moving towards the dam at the bottom of the reservoir and monitoring the relevant parameters of the density current.

[0059] The described reservoir discharge mode adjustment includes adjusting the outflow and the position of the discharge outlet.

[0060] The relevant parameters of the headward erosion and density current sediment flushing include the sediment movement velocity at the reservoir bottom, the change rate of the water level in front of the dam, the shape, velocity, and sediment concentration of the density current.

[0061] Step S5: During the collaborative sediment flushing scheduling process, monitor the operation indicators of the reservoir group and multiple downstream targets in real time, achieve dynamic balance of each target based on multi-objective optimization, and adjust the scheduling plan in real time according to the monitoring situation; In step S5, the multi-objective optimization is achieved based on a multi-objective optimization algorithm. The multi-objectives include flood control, sediment reduction, ecology, and water resource utilization. The weighted sum method is used in the multi-objective optimization algorithm, and the formula is as follows: ; Among them, is the comprehensive objective function value, is the th objective function, is the weight of the th objective function, is the decision variable vector, is the number of objective functions.

[0062] In a specific implementation, in the cascade reservoir joint scheduling, there are three objectives: maximizing the flood control storage , minimizing the sediment deposition , and maximizing the power generation benefit . The decision variable vector , where represents the reservoir discharge in different time periods. Assuming that through analysis and modeling, the flood control storage objective function ( is the duration of each time period), the sediment deposition objective function ( is the incoming sediment concentration in each time period), and the power generation benefit objective function is the head related to the discharge , is the density of water, is the acceleration due to gravity, is the power generation efficiency). And according to the key points of reservoir management, determine the flood control objective weight , the sediment reduction objective weight , and the power generation objective weight .

[0063] Specifically: Step S501: The core idea of the weighted sum method is to combine multiple objective functions into a single comprehensive objective function through weighting, thus transforming the multi-objective optimization problem into a single-objective optimization problem.

[0064] Step S502: First, clarify the meaning and calculation method of each objective function. The flood control storage capacity objective function indicates that as the discharge increases, the flood control storage capacity will decrease accordingly. The sediment deposition amount objective function is proportional to the discharge and the sediment concentration in the incoming water. The power generation benefit objective function is related to the discharge, water head, and power generation efficiency.

[0065] Step S503: Then, according to the weighted sum method formula , multiply each objective function by its corresponding weight and then sum them up.

[0066] Step S504: The weighted value of the flood control objective is .

[0067] Step S505: The weighted value of the sediment reduction objective is .

[0068] Step S506: The weighted value of the power generation objective is .

[0069] Step S507: Comprehensive objective function: .

[0070] Step S508: Next, use the gradient descent method to solve for the decision variable vector that maximizes .

[0071] Step S6: After the scheduling is completed, comprehensively evaluate the scheduling effect, and feedback the evaluation results to the above dynamic model and scheduling start condition setting link for learning and optimization; In Step S6, the comprehensive evaluation indicators include changes in sediment deposition amount, improvement in flood control ability, improvement in ecological environment, changes in shipping conditions, and water resource utilization efficiency. The learning and optimization use a machine learning algorithm to process the feedback data. The machine learning algorithm uses a linear regression model, and the formula is as follows: ; where, is the predicted value, is the intercept, is the regression coefficient, is the independent variable.

[0072] It should be noted that, in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or apparatus comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or elements inherent to such process, method, article or apparatus.

[0073] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An optimized sediment regulation method for reservoirs during the flood season under the joint operation of cascade reservoirs, characterized in that It includes the following steps: Step S1: Collect multi-source data related to sediment regulation during the flood season of the reservoir, perform fusion processing and analysis on the collected multi-source data, and construct a dynamic model reflecting the current water and sediment conditions and the interaction relationship between factors. Step S2: Based on the dynamic model in Step S1 and historical multi-source data, adopt an adaptive optimization method to adjust the scheduling start conditions according to different water and sediment characteristics. Step S3: Before making a scheduling decision, conduct pre-simulations on different scheduling plans, perform risk assessments on the pre-simulation plans based on the mutual influence and complex conditions of the reservoir group, and screen out a set of plans whose comprehensive benefits meet the requirements. Step S4: According to the real-time monitoring and pre-simulation results, initiate the collaborative scheduling of headward erosion and density current sediment discharge, and monitor and adjust the relevant parameters of headward erosion and density current sediment discharge during the scheduling process. Step S5: During the collaborative sediment discharge scheduling process, monitor the operation indicators of the reservoir group and multiple downstream targets in real time, achieve dynamic balance of each target based on multi-objective optimization, and adjust the scheduling plan in real time according to the monitoring situation. Step S6: After the scheduling ends, conduct a comprehensive evaluation of the scheduling effect, and feedback the evaluation results to the above dynamic model and the scheduling start condition setting link for learning and optimization.

2. The optimized sediment regulation method for the flood season of the lower reservoir under the cascade reservoir joint regulation according to claim 1, wherein: In Step S1, the multi-source data includes reservoir hydrological data, meteorological data, basin land use change data, and reservoir surrounding topography and geomorphology data; the weighted average fusion function is used to perform fusion processing on the multi-source data, and the formula is as follows: ； Among them, is the data value after fusion, is the data value of the th data source, is the weight corresponding to the th data source, and , is the number of types of data sources.

3. A method for optimizing sediment regulation during the flood season of a downstream reservoir under the joint operation of cascade reservoirs according to claim 2, characterized in that: The reservoir hydrological data includes the flow rate and sediment concentration data of the main control stations for inflow and outflow of multiple reservoirs in the upper and lower reaches of the river or river section and other hydrological stations in the reservoir area.

4. An optimized sediment regulation method for the flood season of a downstream reservoir under the joint operation of cascade reservoirs according to claim 1, characterized in that: In Step S2, the adaptive optimization method adopts a dynamic programming algorithm combined with a machine learning algorithm. The dynamic programming is used to adjust the scheduling start conditions according to different water and sediment characteristics. The one-dimensional dynamic programming recurrence formula is as follows: ； Among them, is the optimal value function at stage when the state is . is the decision variable at stage . is the set of allowable decisions in state . is the immediate reward obtained by taking decision from state at stage . is the discount factor, is the state of the next stage determined by state and decision .

5. The optimal sediment regulation method for the flood season of the lower reservoir under the joint operation of cascade reservoirs according to claim 1, characterized in that: In Step S3, the complex conditions include extreme water and sediment conditions. When performing a risk assessment on the pre-simulation plan, a linear weighted combination method is used to evaluate multiple risk indicators, and the formula is as follows: ； Among them, is the comprehensive risk value, is the th risk index value, is the th weight of the risk index, is the number of risk indices.

6. The optimized sediment regulation method for the flood season of the lower reservoir under the joint operation of cascade reservoirs according to claim 1, characterized in that: In Step S4, to initiate headward erosion, the upstream reservoir is controlled to adjust the water level in front of the dam to form a drop with a height difference from the upstream silt apex, and the movement of sediment at the bottom of the reservoir and the change of the water level in front of the dam are monitored. For density current sediment discharge, the reservoir discharge mode is adjusted to induce the formation of a stable density current moving towards the dam at the bottom of the reservoir, and the relevant parameters of the density current are monitored.

7. An optimized sediment regulation method for the flood season of a downstream reservoir under the joint operation of cascade reservoirs according to claim 6, characterized in that: Adjusting the reservoir discharge mode includes adjusting the outflow rate and the position of the discharge outlet.

8. A method for optimizing sediment regulation during the flood season of a reservoir under cascade reservoir joint operation according to claim 6, characterized in that: The relevant parameters of headward erosion and density current sediment discharge include the movement speed of sediment at the bottom of the reservoir, the change rate of the water level in front of the dam, the shape, speed, and sediment concentration of the density current.

9. The optimal sediment regulation method for the downstream reservoir during the flood season under the joint operation of cascade reservoirs according to claim 1, characterized in that: In Step S5, multi-objective optimization is achieved based on a multi-objective optimization algorithm. The multi-objectives include flood control, sediment reduction, ecology, and water resource utilization. The weighted sum method is used in the multi-objective optimization algorithm, and the formula is as follows: ； Among them, is the comprehensive objective function value, is the th objective function, is the th weight of the objective function, is the decision variable vector, is the number of objective functions.

10. A method for optimizing sediment regulation during the flood season of a reservoir under the combined operation of cascade reservoirs according to claim 1, characterized in that: In the step S6, the indicators for comprehensive evaluation include changes in sediment deposition volume, improvement in flood control capacity, improvement in ecological environment, changes in shipping conditions, and water resource utilization efficiency. The learning and optimization process uses a machine learning algorithm to process the feedback data. The machine learning algorithm uses a linear regression model, and the formula is as follows: ； Among them, is the predicted value, is the intercept, is the regression coefficient, is the independent variable.

Images