Intelligent water conservancy-oriented drainage basin water engineering control application optimization regulation and control system and method

Abstract

The invention discloses an intelligent water conservancy-oriented drainage basin water engineering control application optimization regulation and control system and method. The regulation and control system comprises a watershed topology intelligent processor, a reservoir gate dam control plan fine simulator, a reservoir gate dam scheduling plan simulator, a flood discharge storage stagnation space water flow simulator, a scheduling plan multi-target evaluator and an engineering regulation and control intelligent optimizer, the method is used for intelligently constructing an optimal regulation and control project, an upstream control application project and a downstream control application project set of a drainage basin reservoir project group and a gate dam project group. According to the invention, by introducing the basin topology intelligent processor, intelligent construction of the basin water engineering partition set and the dominating relationship is realized, and a foundation is laid for parallel computing; a multi-level and high-precision water engineering and water flow process simulation system is constructed by combining a reservoir gate dam control plan fine simulator, a scheduling plan simulator and a flood discharge and storage stagnation space water flow simulator.

Description

Technical Field

[0001] The present invention relates to an optimized regulation system and method for the control and operation of basin water projects for intelligent water conservancy, belonging to the fields of water conservancy and hydropower engineering and water environmental protection. Background Art

[0002] In basin management, the joint regulation of water project groups such as reservoirs and sluices is considered the core means to improve comprehensive benefits such as flood control, water supply, and power generation. However, due to the complex decision-making process of the optimized regulation of basin water project groups involving multiple objectives (such as flood control safety and water supply guarantee) and multiple constraints (such as water volume balance and project storage capacity limitation), traditional scheduling methods often show limitations of low computational efficiency and poor optimization effect when facing large-scale, high-dimensional, and strongly non-linear problems. These problems make it difficult for traditional methods to meet the requirements of rapid response and precise decision-making in modern basin management.

[0003] The preview and pre-plan functions need to efficiently simulate and optimize the regulation plans of basin water project groups in a short time to cope with complex and changeable water regime changes. However, there are still deficiencies in several key aspects of the existing technologies. For example, the processing of the topological relationship of water project groups in the basin relies on manual analysis, lacking intelligent and automated means to cope with the influence of recent human activities in the basin; the refined simulation ability is limited and cannot fully reflect the dynamic interaction between water flow and project regulation; in addition, in terms of computational efficiency, there is a lack of effective parallel computing support, resulting in too long time-consuming for the optimization of large-scale regulation plans. These technical bottlenecks limit the rapid response ability of basin water project groups under extreme hydrological events and the full play of comprehensive benefits. Summary of the Invention

[0004] Object of the Invention: To overcome the deficiencies in the existing technologies, the present invention provides an optimized regulation system and method for the control and operation of basin water projects for intelligent water conservancy. By introducing a basin topology intelligent processor, the intelligent construction of the partition set and dominance relationship of basin water projects is realized, laying a foundation for parallel computing; combined with a fine simulator for reservoir and sluice control pre-plans, a scheduling pre-plan simulator, and a water flow simulator for flood detention and storage spaces, a multi-level and high-precision simulation system of water projects and water flow processes is constructed.

[0005] Technical Solution: To solve the above technical problems, the optimized regulation system for the control and operation of basin water projects for intelligent water conservancy of the present invention includes a basin topology intelligent processor, a fine simulator for reservoir and sluice control pre-plans, a reservoir and sluice scheduling pre-plan simulator, a water flow simulator for flood detention and storage spaces, a multi-objective evaluator for scheduling pre-plans, and an intelligent optimizer for project regulation; The described basin topology intelligent processor is used to intelligently construct the neighborhood domination matrix D1 and the global domination matrix D2 of the reservoir and sluice project group in the basin, analyze and obtain the upstream control operation project set S1, the optimized regulation project set S2, the downstream control operation project set S3, and the domination camps within the sets, and support the parallel computing of the simulation analysis of the basin project group; it is used to obtain the specified scheduling methods and parameters of each reservoir and sluice project in the basin, and obtain the inflow of each project and its evolution lag time statement. The described reservoir and sluice control plan fine simulator calculates the synthetic inflow process of the project according to the specified calculation of the inflow of the control operation project and its evolution lag time statement. The described reservoir and sluice scheduling plan simulator participates in the calculation of the synthetic inflow process of the project according to the specified calculation of the inflow of the optimized regulation project and its evolution lag time statement. The described flood routing and storage space water flow simulator identifies the most downstream project in the basin reservoir and sluice projects that does not dominate other projects at all through the global domination matrix D2. Facing the downstream flood routing and storage space of the most downstream project, with the outflow calculated by the most downstream project as the upstream incoming water boundary, based on the river channel confluence hydrology method or the hydrodynamic method, considering the influence of flood diversion in the flood storage and detention area, pumping drainage by pumping stations, and inter-basin confluence factors, it conducts the time-step simulation of the continuous water flow calculation in the flood routing and storage space, outputs the hydraulic element process of the control section in the flood routing and storage space within the calculation time domain, and quantitatively determines the violation amount of the flood routing and storage space constraints corresponding to the current upstream incoming water boundary. The involved constraint conditions include: hydraulic element range constraint, hydraulic element amplitude change constraint, hydraulic element rate of change constraint, hydraulic element threshold damage depth constraint, and hydraulic element threshold damage duration constraint. The described scheduling plan multi-objective evaluator conducts multi-objective evaluation for the current optimized regulation project scheduling plan. The described project regulation intelligent optimizer consists of three parts: a plan population initialization unit, a plan population quantitative evaluation unit, and a plan population heuristic evolution unit; among them, the plan population initialization unit randomly generates a two-dimensional matrix. Each row of the matrix represents a scheduling plan for all the projects participating in the optimized regulation, that is, the upstream water level or the downstream discharge process of each project within the calculation time domain. The matrix is formed by combining multiple scheduling plans column by column, which is the plan population; the described plan population quantitative evaluation unit assigns a quantitative priority to each scheduling plan using an adaptive constraint handling method based on the weighted objective deviation amount and the total constraint violation amount of each scheduling plan in the population; the described plan population heuristic evolution unit iteratively updates and evolves the plan population using a heuristic evolution algorithm based on the quantitative priorities of each scheduling plan.

[0006] Preferably, the basin topology intelligent processor constructs a basin project neighborhood domination matrix D1, where D1 is a two-dimensional square matrix with dimensions of M + N, M is the number of reservoir projects, N is the number of sluice and dam projects, and the value of D1ij being 1 represents that the outflow of project i directly flows into the adjacent downstream project j, that is, project i affects and dominates project j. A value of -1 represents that project i directly receives the outflow of the adjacent upstream project j, that is, project i is affected and dominated by project j. A value of 0 represents that there is no mutual influence and domination relationship between project i and project j. Based on the neighborhood domination matrix D1, matrix analysis is carried out to obtain the global domination matrix D2. D2 is a two-dimensional square matrix with dimensions of M + N. The value of D2ij being 1 represents that the outflow of project i directly or indirectly flows into the downstream project j, that is, project i affects and dominates project j. A value of -1 represents that project i directly or indirectly receives the outflow of the upstream project j, that is, project i is affected and dominated by project j. A value of 0 represents that there is no mutual influence and domination relationship between project i and project j.

[0007] A regulation method for the above-mentioned basin water project control operation optimization regulation system for intelligent water conservancy includes the following steps: Step 1: Apply the basin topology intelligent processor to obtain the designated scheduling methods and parameters of each reservoir and sluice and dam project in the basin, the inflow of each project and its evolution lag time declaration, construct the basin project neighborhood domination matrix D1 and the global domination matrix D2, divide the basin reservoir and sluice and dam group into S1, S2, and S3. S1 is the set of upstream control operation projects, S2 is the set of optimization regulation projects, and S3 is the set of downstream control operation projects. Use the fast non-dominated sorting algorithm to obtain multiple domination camps within each project set. Among them, the outflow of the projects contained in the upper camp is used as the inflow of the projects contained in the lower camp. For all projects belonging to the same camp, there is no intersection between their outflows and inflows. Record the CPU time at the start of Step 1; Step 2: Based on the reservoir and sluice and dam control plan fine simulator, for the set of upstream control operation projects S1, carry out engineering simulation calculations for each domination camp one by one: First, calculate the current project's combined inflow process according to the inflow and its evolution lag time declaration. Then, apply the mathematical models of various control operation modes of different types of projects, and combine the current project's control parameters to perform continuous time-step simulations, and output the upstream water level of the current project and the calculated outflow process within the calculation time domain; Step 3: Based on the engineering regulation intelligent optimizer, use the pre-plan population initialization unit to randomly generate a pre-plan population, that is, a two-dimensional matrix where each row represents a scheduling pre-plan for all projects participating in the optimization regulation. The scheduling pre-plan is the upstream water level or the downstream discharge process of the project within the calculation time domain; Step 4: Construct an empty historical pre-plan library to store the evaluated optimization regulation project scheduling pre-plans, as well as the corresponding reservoir and sluice and dam project and flood storage and detention space simulation results, weighted target deviation amounts, and total constraint violation amounts; Step 5: For each scheduling plan of the optimized regulation project in the current plan population, if an evaluated scheduling plan can be located in the historical plan library, directly retrieve the simulation results of the reservoir sluice project and the flood storage and detention space, the weighted target deviation, and the total constraint violation; otherwise, conduct simulation calculations to obtain the simulation results of the reservoir sluice project and the flood storage and detention space, the weighted target deviation, and the total constraint violation, and add them to the historical plan library; Step 6: Use the plan population quantitative evaluation unit of the project regulation intelligent optimizer to assign a quantitative priority to each scheduling plan based on the weighted target deviation and the total constraint violation of each scheduling plan in the population using an adaptive constraint handling method; The specific calculation method is as follows: ; Where, is the priority of a scheduling plan in the population, is the weighted target deviation of this scheduling plan, is the total constraint violation of this scheduling plan, is the proportion of feasible plans with a total constraint violation of 0 in the population, is the deviation of each scheduling target of the optimized regulation project and the flood storage and detention space is the weight coefficient, is the violation of each constraint of the optimized regulation project and the flood storage and detention space; Step 7: Use the plan population heuristic evolution unit of the project regulation intelligent optimizer to update the scheduling plan population of the optimized regulation project using a heuristic evolution algorithm based on the quantitative priorities of each scheduling plan, and read the CPU time at this moment ; Repeat steps 5 to 7 until the exit condition is met. The exit conditions include that the total calculation time reaches the preset upper limit, or one or more of the iteration update times reach the preset upper limit, and calculate the total calculation time : ; The calculation method of the iteration update times is: the number of repeated executions of steps 5 to 7; after exiting, output the scheduling plan of the optimized regulation project with the highest priority during the iterative calculation.

[0008] Preferably, the method of applying the fast non-dominated sorting algorithm to matrices D1 and D2 is as follows: A11. If no project in the basin participates in the optimized regulation project, all projects belong to the upstream control operation project set S1; if at least one project in the basin participates in the optimized regulation project, the optimized regulation project set S2 is not empty; if there is a control operation project within the scope of two optimized regulation projects, the participating optimized regulation project collection is a subset of the optimized regulation project set S2; A12. If the set S2 of optimized regulation projects is not empty, starting from all the projects participating in the optimized regulation projects, continuously locate the adjacent downstream projects according to the matrices D1 and D2 and store them in the union of S2 and S3. After obtaining the union of S2 and S3, use the element exclusion method to obtain the set S1 of upstream control operation projects; A13. According to the neighborhood domination matrix D1 and the global domination matrix D2, determine the reservoir and sluice projects that do not dominate other projects at all, do not belong to the set S1 of upstream control operation projects, and are designated to carry out control operations. Starting from the reservoir and sluice projects, continuously locate the adjacent upstream projects and store them in the set S3 of downstream control operation projects. The condition that the adjacent upstream projects should meet is that they do not belong to the set S1 of upstream control operation projects and are designated to carry out control operations. If there are no adjacent upstream projects that meet the conditions during the location process, the location ends; A14. On the basis of obtaining the set S1 of upstream control operation projects and the set S3 of downstream control operation projects, use the element exclusion method to obtain the set S2 of optimized regulation projects; A15. For the sets S1, S2, and S3 respectively, extract the corresponding rows and columns of the matrices D1 and D2 as new temporary domination matrices, and apply the fast non-dominated sorting algorithm to divide each project set into different domination camps.

[0009] Preferably, in step 5, the simulation calculation includes the following steps: Step 5.1: For each domination camp in the set S2 of optimized regulation projects, carry out project type judgment and simulation calculation in the following way: If the current project is a control operation project, based on the refined simulator of the reservoir and sluice control plan, calculate the synthetic inflow process of the current project according to the inflow and its evolution lag time statement, apply the mathematical models of various control operation modes of different types of projects, and perform continuous calculation time step simulation in combination with the control parameters of the current project, and output the upstream water level and the calculated outflow process of the current project within the calculation time domain; If the current project is an optimized regulation project, based on the reservoir and sluice operation plan simulator, calculate the synthetic inflow process of the current project according to the inflow and its evolution lag time statement, based on the operation plan of the current project, perform continuous calculation time step simulation in combination with the optimization parameters, and output the upstream water level, the calculated inflow, and the calculated outflow process of the current project within the calculation time domain, and quantify the constraint violation amount of each project at each calculation time step; Step 5.2: Based on the refined simulator of the reservoir and sluice control plan, for the set S3 of downstream control operation projects, carry out simulation calculation for each project in each domination camp: First, calculate the synthetic inflow process of the current project according to the inflow and its evolution lag time statement, and then apply the mathematical models of various control operation modes of different types of projects, and perform continuous calculation time step simulation in combination with the control parameters of the current project, and output the upstream water level and the calculated outflow process of the current project within the calculation time domain; Step 5.3: If the constraint violation amounts of each project at each calculation time step in Step 5.1 are all 0, based on the above-mentioned flow-retaining and flood-detention space water flow simulator, with the calculated outflow process of the project at the most downstream of the basin as the upstream incoming water boundary, conduct continuous calculation time step simulation of the flow-retaining and flood-detention space water flow, output the hydraulic element process of the control section of the flow-retaining and flood-detention space within the calculation time domain, and quantitatively determine the constraint violation amount of the flow-retaining and flood-detention space; otherwise, directly assign a very large value to the constraint violation amount of the flow-retaining and flood-detention space. Step 5.4: Based on the above-mentioned multi-objective evaluator for the dispatching plan, conduct multi-objective evaluation for the current optimized regulation project dispatching plan. On the basis of quantitatively determining the constraint violation amount of the optimized regulation project and the constraint violation amount of the flow-retaining and flood-detention space, quantitatively determine the deviation of each dispatching objective of the optimized regulation project and the flow-retaining and flood-detention space, and then comprehensively determine the weighted objective deviation amount and the total constraint violation amount. Step 5.5: Add the simulation results of the reservoir sluice projects and the flow-retaining and flood-detention space, the weighted objective deviation amount and the total constraint violation amount of the current simulated evaluation dispatching plan to the historical plan library.

[0010] Beneficial effects: Compared with the prior art, the present invention has the following advantages: (1) Intelligent topology processing: Intelligently divide the project set and the dominant camp through the dominance matrix, improve the efficiency and accuracy of topology processing, and lay a foundation for parallel simulation calculation; (2) Fine and efficient simulation: Use various control mode mathematical models to accurately simulate the change of project water level and flow rate and the downstream flow process; (3) Intelligent optimization and rapid response: Combine adaptive constraint processing and heuristic algorithm, quickly search for and optimize the preferred plan, and accelerate the response through parallel calculation. Brief description of the drawings

[0011] Figure 1 is the structural diagram of the optimized regulation system for basin water project control application facing intelligent water conservancy of the present invention; Figure 2 is the flow chart of the method for accelerating the optimized regulation calculation of the basin water project group control application of the present invention; Figure 3 is the flow chart of Step 5 of the present invention. Detailed implementation manners

[0012] The present invention will be further described in detail below with reference to the accompanying drawings.

[0013] Embodiment 1: In the present invention, consider a basin containing 10 projects: 6 reservoirs (R1, R2, R3, R4, R5, R6, M = 6) and 4 sluice dams (G1, G2, G3, G4, N = 4), with a total of M + N = 10. The project connection relationship is as follows: The outflow of reservoir R1 flows into R2; The outflow of reservoir R2 flows into R3; The outflows of reservoirs R3 and R4 jointly flow into R5; The outflows of reservoirs R5 and R6 jointly flow into the sluice dam G1; The outflows of sluice dams G1 and G2 jointly flow into the sluice dam G3; The outflows of sluice dams G3 and G4 jointly flow into the downstream river; The dimensions of matrices D1 and D2 are 10×10.

[0014] First, construct the neighborhood domination matrix D1: The neighborhood domination matrix D1 represents the direct influence relationship between projects: D1[i,j]=1 indicates that the outflow of project i directly flows into the adjacent downstream project j, that is, i affects j; D1[i,j]=-1 indicates that i receives the outflow of j, that is, i is affected by j; D1[i,j]=0 indicates that there is no direct influence, and the diagonal element D1[i,i]=0 because a project does not affect itself.

[0015] Analyze the connection relationship: R1 flows into R2: D1[R1,R2]=1, D1[R2,R1]=-1; R2 flows into R3: D1[R2,R3]=1, D1[R3,R2]=-1; R3 and R4 flow into R5: D1[R3,R5]=1, D1[R4,R5]=1, D1[R5,R3]=-1, D1[R5,R4]=-1; R5 and R6 flow into G1: D1[R5,G1]=1, D1[R6,G1]=1, D1[G1,R5]=-1, D1[G1,R6]=-1; G1 and G2 flow into G3: D1[G1,G3]=1, D1[G2,G3]=1, D1[G3,G1]=-1, D1[G3,G2]=-1.

[0016] The neighborhood domination matrix D1 is: ; Subsequently, construct the global domination matrix D2: The global domination matrix D2 reflects the direct or indirect influence relationship through matrix analysis (such as transitive closure) based on D1: D2[i,j]=1 indicates that the outflow of i directly or indirectly flows into j, that is, i affects j; D2[i,j]=-1 indicates that i receives the outflow of j, that is, i is affected by j; D2[i,j]=0 indicates no influence. The analysis is as follows: The influence of R1: directly affects R2, affects R3 through R2, affects R5 through R3, affects G1 through R5, and affects G3 through G1. Therefore, D2[R1, R2]=1, D2[R1, R3]=1, D2[R1, R5]=1, D2[R1, G1]=1, D2[R1, G3]=1; The influence of R2: directly affects R3, affects R5 through R3, affects G1 through R5, and affects G3 through G1. Therefore, D2[R2, R3]=1, D2[R2, R5]=1, D2[R2, G1]=1, D2[R2, G3]=1; The influence of R3: directly affects R5, affects G1 through R5, and affects G3 through G1. Therefore, D2[R3, R5]=1, D2[R3, G1]=1, D2[R3, G3]=1; The influence of R4: directly affects R5, affects G1 through R5, and affects G3 through G1. Therefore, D2[R4, R5]=1, D2[R4, G1]=1, D2[R4, G3]=1; The influence of R5: directly affects G1, and affects G3 through G1. Therefore, D2[R5, G1]=1, D2[R5, G3]=1; The influence of R6: directly affects G1, and affects G3 through G1. Therefore, D2[R6, G1]=1, D2[R6, G3]=1; The influence of G1: directly affects G3. Therefore, D2[G1, G3]=1; The influence of G2: directly affects G3. Therefore, D2[G2, G3]=1; Global domination matrix D2: .

[0017] Subsequently, assuming that reservoirs R5, R6 and sluice dam G1 participate in the optimal operation, and other projects are under control operation, the method of applying the fast non-dominated sorting algorithm to matrices D1 and D2 is further illustrated.

[0018] Among them, the fast non-dominated sorting algorithm is implemented as follows: The fast non-dominated sorting algorithm (Fast Nondominated Sorting Algorithm) is an algorithm derived from multi-objective optimization. The main steps of fast non-dominated sorting are as follows: ① Initialization: For each reservoir and sluice project, initialize two parameters: the number of water projects n(p) that dominate this water project, and the set S(p) of water projects dominated by this water project.

[0019] ②Non - dominated level identification: Identify all water projects with n(p) = 0, that is, these water projects are not dominated by any other water projects. These water projects form a camp of water projects, and mark these water projects as the camp of the current layer (the first layer).

[0020] ③Camp expansion: For each water project p in the current camp, consider the set of water projects S(p) it dominates. For each water project q in S(p), subtract 1 from the domination number n(q) of q. When n(q) is reduced to 0, it means that the water project q is not dominated by other water projects outside the currently identified camp, so q belongs to the next - layer camp. Collect all newly identified non - dominated water projects to form a new non - dominated camp level, and then continue to repeat this process until all water projects are classified into camp levels.

[0021] (1)Judge whether S2 is empty If at least one project participates in the optimal regulation, then S2 is not empty. R5, R6, and G1 are the projects for optimal regulation, so S2 is not empty. Initially, S2 = {R5, R6, G1}. There is no need to consider the situation of "control operation projects appearing within the scope of two optimal regulation projects" (there are no control operation projects between the projects in the initial S2).

[0022] Other projects (R1, R2, R3, R4, G2, G3, G4) are to be allocated.

[0023] (2)Locate the union of S2 and S3 to determine S1 Taking the optimal regulation projects (R5, R6, G1) in S2 as the starting point, locate the adjacent downstream projects according to the matrix D2: R5: D2[R5, G1] = 1, D2[R5, G3] = 1, affecting G1 and G3.

[0024] R6: D2[R6, G1] = 1, D2[R6, G3] = 1, affecting G1 and G3.

[0025] G1: D2[G1, G3] = 1, affecting G3. The downstream project is G3 (G1 is already included in S2), deposit it into the union of S2 and S3, and the current union is {R5, R6, G1, G3}. Using the element exclusion method, obtain S1: The universal set {R1, R2, R3, R4, R5, R6, G1, G2, G3, G4} minus {R5, R6, G1, G3}, getting S1 = {R1, R2, R3, R4, G2, G4}.

[0026] (3)Determine S3 Identify the projects that do not dominate other projects at all according to D2 (D2[i, j]≠1 for all j≠i), and do not belong to S1 and are for control operation: Row G3: D2[G3,j] = 0 (for j ≠ G3), which does not dominate other processes at all.

[0027] Row G4: D2[G4,j] = 0 (for all j), which does not dominate other processes at all.

[0028] For other processes (R1, R2, R3, R4, R5, R6, G1, G2), there are positive values in D2, so they are excluded. Check S1 = {R1, R2, R3, R4, G2, G4}: G4 is in S1, so it is excluded; G3 is not in S1 and is for control operation, which meets the conditions. Starting from G3, locate the adjacent upstream processes (D2[G3,i] = -1, and i is not in S1 and is for control operation): D2[G3,R1] = -1, and R1 is in S1, so it is excluded.

[0029] D2[G3,R2] = -1, and R2 is in S1, so it is excluded.

[0030] D2[G3,R3] = -1, and R3 is in S1, so it is excluded.

[0031] D2[G3,R4] = -1, and R4 is in S1, so it is excluded.

[0032] D2[G3,R5] = -1, and R5 is for optimized regulation, so it is excluded.

[0033] D2[G3,R6] = -1, and R6 is for optimized regulation, so it is excluded.

[0034] D2[G3,G1] = -1, and G1 is for optimized regulation, so it is excluded.

[0035] D2[G3,G2] = -1, and G2 is in S1, so it is excluded. There is no upstream process that meets the conditions, and the location ends. Therefore, S3 = {G3}.

[0036] (4)Determine S2 Based on S1 = {R1, R2, R3, R4, G2, G4} and S3 = {G3}, use the element exclusion method: The universal set {R1, R2, R3, R4, R5, R6, G1, G2, G3, G4} minus S1 ∪ S3 = {R1, R2, R3, R4, G2, G4, G3}, and we get S2 = {R5, R6, G1}, which is consistent with the initial judgment.

[0037] (5)Divide the domination camps Apply the fast non-dominated sorting algorithm to S1, S2, and S3 respectively, based on the corresponding rows and columns of D2: S1 = {R1, R2, R3, R4, G2, G4}: Extract the rows and columns of R1, R2, R3, R4, G2, G4 in D2: Fast non-dominated sorting: Projects that are completely independent of other projects (D2[i,j] ≠ -1 for all j ≠ i): First camp: C1 = {R1, R4, G2, G4} (completely independent of other projects within S1).

[0038] Second camp: C2 = {R2}.

[0039] Third camp: C3 = {R3}.

[0040] S2 = {R5, R6, G1}: Extract the rows and columns of R5, R6, and G1 from D2: Fast non - dominated sorting: R5 and R6 dominate G1 (D2[R5, G1] = D2[R6, G1] = 1), and they do not dominate each other (D2[R5, R6] = D2[R6, R5] = 0). Both are classified into the first camp C1 = {R5, R6}.

[0041] After removing R5 and R6, there is only one project G1, which is classified into the second camp C2 = {G1}.

[0042] The first camp of S2 is C1 = {R5, R6}, and the second camp is C2 = {G1}.

[0043] S3 = {G3}: There is only one project, and the first camp is C1 = {G3}.

[0044] Example 2: This example focuses on a small watershed in the middle and lower reaches of the Yangtze River in China. Watershed background information: This watershed is located in the humid area of southern China, with a watershed area of 8900 km², an average annual precipitation of 1020 mm, the flood season concentrated from May to October, and the dry season from November to April of the following year. The watershed has 4 reservoirs (JJ, DSH, ZH, LYD) and 3 sluice dams (DC, JZ, HS), which together form the water conservancy regulation system of the watershed: The outflow of the leading reservoir JJ converges into reservoir DSH after a 10 - hour evolution lag time; The outflow of reservoir DSH (6 - hour evolution lag time) and the outflow of reservoir ZH (8 - hour evolution lag time) together converge into sluice dam DC; The outflow of sluice dam DC (4 - hour evolution lag time) converges into sluice dam JZ; The outflow of sluice dam JZ (3 - hour evolution lag time) converges into sluice dam HS, and the outflow of reservoir LYD directly converges into sluice dam HS after a 10 - hour evolution lag time; The outflow of sluice dam HS finally enters the downstream river channel.

[0045] I. Watershed Topological Intelligent Processor (I) Collection and arrangement of project information The designated operation modes and parameters of each reservoir, sluice and dam project in the basin. Among them, reservoirs JJ, ZH, and LYD are control operation projects, and the control operation parameter of reservoir JJ is the target control water level of 200 meters; reservoirs DSH and sluice DC are optimized regulation projects, and the expected upstream water level of reservoir DSH is 150 meters. At the same time, obtain the inflow of each project and its evolution lag time statement. For example, the inflow of reservoir DSH includes the outflow of adjacent upstream reservoir JJ and the inflow from the interval. The water flow of reservoir JJ's outflow converges into reservoir DSH after 10 hours.

[0046] (II) Construct the dominance matrix Construct the neighborhood dominance matrix D1. Based on the direct water flow influence between projects, if the outflow of project i directly converges into downstream project j, then D1[i, j]=1; if project i directly receives the outflow of upstream project j, then D1[i, j]=-1; if there is no direct water flow influence relationship between two projects, D1[i, j]=0. For example, from reservoir JJ to reservoir DSH, D1[reservoir JJ, reservoir DSH]=1; reservoir DSH receives the outflow of reservoir JJ, D1[reservoir DSH, reservoir JJ]=-1; there is no direct connection between reservoir JJ and reservoir ZH, D1[reservoir JJ, reservoir ZH]=0.

[0047] Based on the neighborhood dominance matrix D1, construct the global dominance matrix D2 by analyzing the indirect influence relationship. Taking reservoir JJ as an example, its water flow not only directly affects reservoir DSH, but also indirectly affects sluice DC, sluice JZ, and sluice HS through reservoir DSH. Therefore, in D2, D2[reservoir JJ, reservoir DSH]=1, D2[reservoir JJ, sluice DC]=1, D2[reservoir JJ, sluice JZ]=1, D2[reservoir JJ, sluice HS]=1. And so on, determine the values of all elements in D2.

[0048] (III) Project set and dominance camp division Use the fast non-dominated sorting algorithm to conduct set division and camp construction for projects based on matrices D1 and D2. The optimized regulation projects, reservoir DSH and sluice DC, form the optimized regulation project set S2; the upstream control operation projects, reservoirs JJ, ZH, and LYD, form the upstream control operation project set S1; the downstream control operation projects, sluices JZ and HS, form the downstream control operation project set S3.

[0049] Further divide the projects within each set into camps according to the influence relationship. In S1, there is no mutual influence among reservoirs JJ, ZH, and LYD, and they jointly form the first camp: {Reservoir JJ, Reservoir ZH, Reservoir LYD}; in S2, Reservoir DSH affects Dam DC. Reservoir DSH is the first-level camp: {Reservoir DSH}, and Dam DC is the second-level camp: {Dam DC}; in S3, Dam JZ affects Dam HS. Dam JZ is the first-level camp: {Dam JZ}, and Dam HS is the second-level camp: {Dam HS}. In this way, the spatial relationship among the project sets and the influence relationship of the dominant camps are clarified. The outflow of the projects included in the superior camp is declared as the inflow of the projects included in the inferior camp, and there is no intersection between the outflows and inflows of all the projects belonging to the same camp, which is convenient for subsequent parallel simulation calculations in the dominant camps.

[0050] II. Fine Simulator for Reservoir and Dam Control Plan (I) Application of Reservoir Project Control Mode Reservoirs JJ, ZH, and LYD all adopt the pre-dam water level control. Taking Reservoir JJ as an example: Reservoir JJ adopts the pre-dam water level control mode with the goal of maintaining a specific water level (the target water level is 200 meters, and the allowable fluctuation range is 198 - 202 meters). When receiving the confluence from the basin boundary to the current project (the evolution lag time is 0 h), calculate the synthetic inflow process of the project according to the inflow and its evolution lag time. In the subsequent calculation process, carry out the calculation with the water balance equation and adopt the static storage capacity regulation calculation method. Give priority to ensuring the minimum discharge for power generation, and adjust the excess water volume through the flood discharge facilities. By continuously trying different discharge amounts, ensure that the water level at the end of the time period is within the allowable range and meet the requirements such as the upstream water level constraint of the project, the water level range constraints of each water diversion and pumping facility, the discharge capacity range constraints of each water diversion and pumping facility, and the power generation amount range constraints. The initial scheduling water level is 202.3 meters, the incoming flow is 200 cubic meters per second, and after calculation, the discharge is determined to be 220 cubic meters per second (the tail water for power generation is 25 cubic meters per second, and the flood discharge is 195 cubic meters per second), and the water level at the end of the time period reaches 202 meters, meeting the water level requirements.

[0051] (II) Application of Dam Project Control Mode Dams JZ and HS both adopt the upstream water level control. Taking Dam JZ as an example: Dam JZ adopts the upstream water level control mode with the goal of maintaining the upstream water level at about 50 meters (the allowable fluctuation range is 48 - 52 meters) to ensure the domestic, industrial, and irrigation water intake needs of the upstream. If the incoming flow exceeds the maximum discharge capacity of 1400 cubic meters per second, gradually increase the gate opening to ensure that the water level does not exceed 52 meters. In this process, apply the mathematical model of the upstream water level control mode, with the water balance equation as the core, considering the upstream water level constraint of the project, the water level range constraints of the water diversion and pumping facilities, and the discharge capacity range constraints of the water diversion and pumping facilities.

[0052] III. Reservoir Sluice Regulation Plan Simulator 1. Optimization of Regulation Project Parameter Setting For reservoir DSH, the expected upstream water level is the flood limit water level of 150 m, the flood control high water level of 160 m, and the upper limit of the power generation tail water flow is 30 m³ / s; for sluice dam DC, the expected upstream water level is 80 m, the discharge capacity ranges from 1,100 to 1,500 m³ / s, and the upper limit of the water level above the sluice is the design flood water level of 85 m.

[0053] 2. Simulation Calculation Process For reservoir DSH and sluice dam DC, first calculate the synthetic inflow process of the project according to the inflow and its evolution lag time statement. Then, for the specified project participating in the optimization regulation project, based on the regulation plan (in this case, the upstream water level process of reservoir DSH within the calculation time domain and the discharge process of sluice dam DC), combined with the optimization parameters, with the water balance equation as the core, adopt the static storage capacity regulation calculation method to perform continuous time step simulation. During the simulation, output the upstream water level of the project, the calculated inflow, and the calculated outflow process, and quantify the constraint violation amount of each project at each calculation time step, including the upstream water level constraint of the project, the water level range constraint of the water diversion and pumping facilities, the discharge capacity range of the water diversion and pumping facilities, and the power generation range constraint, etc.

[0054] IV. Flow Simulator for Flood Storage and Retention Space 1. Determination of Boundary Conditions Determine that sluice dam HS is the most downstream project through the global domination matrix, and its outflow is used as the upstream incoming water boundary of the downstream river channel.

[0055] 2. Simulation Process Adopt the river channel confluence hydrodynamics method to simulate the water flow in the downstream river channel, consider factors such as interval confluence, and calculate the hydraulic elements such as the water level and flow velocity at the flood control control section CH station of the downstream river channel. During the simulation, set the constraint conditions of the hydraulic elements, such as the water level at the control section cannot exceed the guaranteed water level, and the water level change rate cannot exceed 2 m / h. If the evaluated hydraulic elements exceed these constraint ranges, record the depth and duration of the constraint violation for subsequent evaluation.

[0056] V. Multi-objective Evaluator for Regulation Plan 1. Calculation Dimensions of Target Deviation Optimization of Regulation Project Scheduling Objectives and Deviations: Calculate the average deviation of the water level during the calculation period, that is, the average deviation of the regulated water level from the expected upstream water level parameter in each period; count the proportion of the water level deviation duration, that is, the proportion of the number of periods when the actual water level exceeds the allowable range to the total number of simulation periods; determine the deviation of the final regulated water level of the project, that is, the deviation of the regulated water level at the end of the simulation from the expected upstream water level parameter.

[0057] Downstream river channel regulation objectives and deviations: Calculate the deviation of the hydraulic element objectives at the control section CH station, especially the depth and duration of the simulated water level exceeding the guaranteed water level.

[0058] (2) Priority and weight assignment logic The regulation objectives are ranked in ascending order of priority as the average deviation of the reservoir period, the deviation duration of the reservoir water level, the deviation of the final regulated water level of the reservoir, and the deviation of the river channel flood control safety objective. Through the setting of priorities and weights, the weighted objective deviation and the total constraint violation are comprehensively determined, and a multi-objective evaluation is carried out on the current optimized regulation project scheduling plan.

[0059] VI. Intelligent optimizer for project regulation (1) Details of initializing the plan population Randomly generate a large number of scheduling plans. For the reservoir DSH, the water level ranges from 140 to 160 meters, and a discrete reservoir scheduling water level process is generated at a step of 1 meter; for the sluice dam DC, the discharge ranges from 1100 to 1500 cubic meters per second, and a discrete downstream discharge process is generated at a step of 50 cubic meters per second. Combine these discrete value processes to generate a basic plan, and then randomly perturb 10% of the time step values in the basic plan to finally form 500 initial plans. Each plan is stored in the form of a two-dimensional matrix, with rows representing different plans and columns representing the time course of the optimized regulation project.

[0060] (2) Adaptive constraint handling stage In the initial iteration stage, since there are many plans with constraint violations, in accordance with the principle of the adaptive constraint handling method, plans with small constraint violations are preferentially retained. Even if the objective deviations of these plans are large, they are given a higher selection probability. In the middle and late stages of iteration, when the number of plans that meet the constraints increases, switch to the "objective priority" strategy, calculate the weighted objective deviation of each plan, retain the top 50% of the plans, and maintain the diversity of the plans through crowding degree calculation to prevent falling into local optimal solutions.

[0061] (3) Heuristic evolutionary operations Evolve the plans through crossover and mutation operations. The crossover operation randomly selects two high-quality plans and exchanges the parameters in the middle period (such as exchanging the water level values of the reservoir DSH from the 2nd to the 4th hour) to generate a new plan, so that the new plan inherits the advantages of both sides. The mutation operation fine-tunes the parameters of a certain period of a single project with a probability of 5% (for example, increasing or decreasing the discharge of the sluice dam DC by 10 cubic meters per second in a certain hour) to explore the uncovered parameter space. When the total calculation time exceeds 0.5 hours (the preset upper limit), stop the iteration and output the optimal scheduling plan.

[0062] VII. Calculation acceleration method (1) Optimization of parallel computing strategy Parallel computing is carried out using hardware resources. The upstream controls, namely the reservoirs JJ, ZH, and LYD in the engineering set S1, have no mutual influence and jointly form the first camp. Therefore, for the reservoirs JJ, ZH, and LYD in S1, 3 CPU cores are allocated for separate processing, and each project independently runs a fine simulator for the control plan. As a result, the single-project simulation time is reduced from 3 times in serial to 1 time in parallel. For the reservoirs DSH and the sluice dam DC in the optimized regulation engineering set S2, the parallel computing ability of the GPU is used to simultaneously simulate the scheduling plan, and the single-plan calculation time is reduced from 2 times to 1 time. There is no domination relationship between projects in the same camp, and they can be calculated independently. The upper and lower hierarchical camps are executed in the order of "calculate the upstream level first and the downstream level later" to ensure the correctness of the water flow evolution logic.

[0063] (2) Historical plan library mechanism Before starting iterative optimization, a historical plan library is established to store information such as plan numbers, generation times, optimized regulation engineering parameters, simulation results (including the outputs of upstream control operation projects, downstream control operation projects, and downstream river channel simulation results), and evaluation results (weighted target deviation, total constraint violation, etc.). The plan parameters are encoded using the hash algorithm, and fast query is achieved through the Redis database. As the iteration progresses, the hit rate gradually increases, saving approximately 40% of the simulation time in total, and effectively using the historical plan library to accelerate the calculation.

[0064] At the same time, for the new scheduling plan for which simulation evaluation is carried out, the simulation evaluation results (including target deviation and total constraint violation) are added to the historical plan library.

[0065] (3) Rule for avoiding invalid calculations When a constraint violation occurs during the simulation of the optimized regulation project (such as the water level of reservoir DSH exceeding 160 meters, or the discharge of sluice dam DC exceeding 1500 cubic meters per second), the midstream and downstream simulations of the current plan are no longer carried out (skipping the simulations of sluice dams JZ and HS and the downstream river channel), and the downstream river channel constraint violation amount is directly assigned as a preset maximum value (such as 1000), and the reason for invalidity is recorded to avoid the repeated generation of similar invalid plans, saving a large amount of calculation time and effectively avoiding invalid calculations.

[0066] As Figures 1 to 3 shown, a regulation method for a basin water project control operation optimized regulation system for intelligent water conservancy includes the following steps: Step 1: Apply the described watershed topology intelligent processor to obtain the specified scheduling methods and parameters of each reservoir and sluice project in the watershed, the inflow of each project and its evolution lag time statement, construct the neighborhood domination matrix D1 and the global domination matrix D2 of the watershed projects, divide the reservoir sluice group in the watershed into the upstream control operation project set S1, the optimized regulation project set S2, and the downstream control operation project set S3, and use the fast non-dominated sorting algorithm to obtain multiple domination camps within each project set, where the outflow of the projects contained in the upper camp is used as the inflow of the projects contained in the lower camp, and there is no intersection between the outflow and inflow of all projects belonging to the same camp; The neighborhood domination matrix D1 is as follows: ; The global domination matrix D2 is as follows: ; Upstream control operation project set S1: Reservoirs JJ, ZH, and LYD under upstream control operation; Optimized regulation project set S2: Reservoir DSH and sluice DC participating in optimized regulation; Downstream control operation project set S3: Sluices JZ and HS under downstream control operation; Step 2: Based on the described reservoir and sluice control plan fine simulator, for the upstream control operation project set S1 (reservoirs JJ, ZH, LYD), carry out engineering simulation calculations for each domination camp one by one: First, calculate the current project's combined inflow process according to the inflow and its evolution lag time statement, and then apply the mathematical model of the reservoir project's water level control operation mode in front of the dam, and carry out continuous calculation time step simulation in combination with the current project's control parameters, and output the upstream water level and calculated outflow process of the current project within the calculation time domain; After the simulation of the upstream control operation project set S1 is completed, it will not be recalculated in subsequent iterative optimizations as a measure to accelerate the calculation; Carry out parallel calculations for the project groups within each domination camp as a measure to accelerate the calculation; Step 3: Based on the described engineering regulation intelligent optimizer, use the plan population initialization unit to randomly generate a plan population, that is, a two-dimensional matrix where each row represents a scheduling plan for all projects participating in optimized regulation, and the scheduling plan is the upstream water level or the downstream discharge process of the project within the calculation time domain; For reservoir DSH, within the water level range of 140 - 160 meters, generate a discrete reservoir scheduling water level process at a 1-meter step; For sluice DC, within the discharge range of 1100 - 1500 cubic meters per second, generate a discrete downstream discharge process at a 50-cubic-meter-per-second step. Combine these discrete value processes to generate a basic plan, and then randomly perturb 10% of the time step values in the basic plan to finally form 500 initial plans. Each plan is stored in the form of a two-dimensional matrix, with rows representing different plans and columns representing the time course of the optimized regulation projects.

[0067] Step 4: Construct an empty historical pre - plan library to store the evaluated optimized regulation project scheduling pre - plans, as well as the corresponding reservoir and dam project and flood - routing and storage space simulation results, weighted target deviation, and total constraint violation; Step 5: For each scheduling pre - plan of the optimized regulation project in the current pre - plan population, if an evaluated scheduling pre - plan can be located in the historical pre - plan library, directly retrieve the reservoir and dam project and flood - routing and storage space simulation results, weighted target deviation, and total constraint violation; otherwise, conduct simulation calculations to obtain the reservoir and dam project and flood - routing and storage space simulation results, weighted target deviation, and total constraint violation: The simulation calculations include the following steps: Step 5.1: All projects within the optimized regulation project set S2 are optimized regulation projects (reservoir DSH and dam DC). Each dominant camp conducts project type judgment and simulation calculations as follows: For optimized regulation projects, based on the reservoir and dam scheduling pre - plan simulator, calculate the current project's combined inflow process according to the inflow and its evolution lag time statement. Based on the current project's scheduling pre - plan and combined with optimization parameters, conduct continuous time - step simulations, and output the upstream water level, calculated inflow, and calculated outflow process of the current project within the calculation time domain, and quantify the constraint violation of each project at each calculation time step; Step 5.2: Based on the refined simulator of the reservoir and dam control pre - plan, for the downstream control operation project set S3 (dams JZ, HS), each dominant camp conducts simulation calculations for each project: First, calculate the current project's combined inflow process according to the inflow and its evolution lag time statement, and then apply the mathematical model of the water - level control operation mode above the dam of the dam project, and conduct continuous time - step simulations in combination with the current project's control parameters, and output the upstream water level and calculated outflow process of the current project within the calculation time domain; Step 5.3: If the constraint violation of each project at each calculation time step in Step 5.1 is 0, based on the flood - routing and storage space water flow simulator, use the calculated outflow process of the most downstream project in the basin as the upstream inflow boundary, and conduct continuous time - step simulations of the flood - routing and storage space water flow, and output the hydraulic element process of the control section of the flood - routing and storage space within the calculation time domain, and quantitatively determine the constraint violation of the flood - routing and storage space; otherwise, directly assign a very large value of 1000 to the constraint violation of the flood - routing and storage space; Step 5.4: Based on the multi - objective evaluator of the scheduling pre - plan, conduct multi - objective evaluation for the current optimized regulation project scheduling pre - plan. On the basis of quantitatively determining the constraint violation of the optimized regulation project and the constraint violation of the flood - routing and storage space, quantitatively determine the deviation of each scheduling target of the optimized regulation project and the flood - routing and storage space, and then comprehensively determine the weighted target deviation and the total constraint violation; Step 5.5: Add the reservoir and dam project and flood - routing and storage space simulation results, weighted target deviation, and total constraint violation of the current simulated and evaluated scheduling pre - plan to the historical pre - plan library.

[0068] Step 6: Using the plan population quantitative evaluation unit of the engineering control intelligent optimizer, based on the weighted target deviation and total constraint violation of each scheduling plan in the population, an adaptive constraint processing method is used to assign a quantitative priority to each scheduling plan. The specific calculation method is as follows: ; in, is the priority of a scheduling plan in the population, The weighted target deviation for this plan is: is the total constraint violation of the plan, is the proportion of feasible plans with a total constraint violation of 0 in the population; for a scheduling plan, To optimize the control project, flood storage space and various scheduling target deviations The weight coefficient of To optimize the violation of various constraints of the control project and flood storage space. , so that when the number of plans with a total constraint violation of 0 in the population is small, plans with a smaller total constraint violation are preferred; when the number of plans with a total constraint violation of 0 in the population is large, plans with a smaller weighted target deviation are preferred.

[0069] Step 7: Adopt the plan population heuristic evolution unit of the engineering control intelligent optimizer, and use the heuristic evolutionary algorithm to update the optimized control engineering scheduling plan population based on the quantitative priority of each scheduling plan.

[0070] The quantitative priority of the scheduling plan calculated by the plan population evolution unit based on step 6 P , the heuristic evolutionary algorithm is used to update the population, which is implemented through three core operators of genetic algorithm selection, crossover, and mutation: First, the selection operator adopts the genetic algorithm roulette strategy, according to the priority of each scheduling plan P Calculate the probability of being selected (priority P The higher the probability, the greater the probability of being selected), so as to screen out parent individuals from the current population and provide basic samples for subsequent operations; then, the genetic algorithm crossover operator is used to perform single-point crossover on the selected parent individuals, that is, a decision variable position is randomly selected as the crossover point, and the gene fragments of the two parent individuals after this point are exchanged to generate offspring individuals carrying parental characteristics, thereby inheriting the advantages of the parents and introducing new combination possibilities; finally, the genetic algorithm mutation operator is used to perturb the gene position of the generated offspring individuals with a lower probability (0.04). For each decision variable, if the random number is less than the mutation probability, the variable value is randomly adjusted to increase the population diversity with small changes to avoid falling into the local optimum; the entire evolutionary process updates the scheduling plan population by performing selection, crossover, and mutation operations, so that the population is prioritized.P Gradually evolve towards a better scheduling plan under guidance.

[0071] Repeat steps 5 to 7 until the exit condition is met: when the total calculation time exceeds 0.5 hours (the preset upper limit), stop the iteration and output the optimal scheduling plan. The total calculation time is calculated as follows: before step 7 is executed and jumps to step 5, calculate the total calculation time according to the following formula : , where is the CPU time when step 7 is executed, is the CPU time when step 1 starts to execute. After exiting, output the optimization and control engineering scheduling plan with the highest priority P during the iterative calculation process ( P the calculation method is shown in step 6), that is, the optimal control plan.

[0072] In the present invention, the fine simulators of reservoir sluice control plans, the simulators of reservoir sluice scheduling plans, and the flow simulators in flood storage and detention spaces have exactly the same calculation time domain, and the calculation time steps are determined according to the simulation requirements of reservoir sluice projects and flood storage and detention spaces, and can be different.

[0073] In the present invention, an implementation of the mathematical model of the pre-dam water level control operation mode of the reservoir project is as follows: (1) Select the maximum discharge capacity curve. If the initial water level at the scheduling time period exceeds the maximum controlled discharge water level in the project scheduling regulations, select the maximum sum of the flow capacities of all spillways (open discharge line), otherwise determine the maximum controlled discharge flow curve (controlled discharge line) corresponding to different water levels according to the regulations. Secondly, select the minimum discharge capacity curve: if the maximum discharge capacity curve is the open discharge line, the minimum discharge capacity curve is also the open discharge line, otherwise determine the minimum discharge capacity curve according to the regulations; (2) Perform time interpolation according to the specified target water level control parameters to obtain the target control water level at the end of the current time period. Based on the pre-dam water level at the beginning of the reservoir time period, judge the highest and lowest pre-dam water levels at the end of the time period through the corresponding minimum and maximum discharge capacities, and obtain the reasonable range of the pre-dam water level at the end of the time period. Select the value closest to the target water level at the end of the time period within the reasonable range as the water level at the end of the time period; (3) When the initial water level of the time period is lower than the spillway crest elevation / bottom slab elevation of the sluice, give priority to water storage, and the upper and lower limits of the water level at the end of the time period are the same; (4) When the initial water level of the time period is greater than or equal to the flood limit water level and less than the higher value of the flood limit water level / normal storage water level, consider the relationship between the incoming water in the current time period and the maximum discharge capacity corresponding to the initial water level. If the incoming water is large, iteratively calculate a higher water level at the end of the time period, otherwise iteratively calculate a lower water level at the end of the time period, and thus obtain the lower limit of the water level at the end of the time period; when all the incoming water is stored without discharge, obtain the water level at the end of the time period, and thus obtain the upper limit of the water level at the end of the time period; (5) If the initial water level at the beginning of the time period is greater than or equal to the higher value of the flood limit water level / normal storage water level, considering the relationship between the incoming water in the current time period and the maximum discharge capacity corresponding to the initial water level, if the incoming water is large, iteratively calculate a higher water level at the end of the time period, and vice versa, iteratively calculate a lower water level at the end of the time period, thereby obtaining the lower limit of the water level at the end of the time period; for the upper limit of the water level at the end of the time period, take the trial value of the water level at the end of the time period. (6) For the calculated time period discharge, if the initial and end water levels of the time period both meet the constraints of the power generation diversion water level range, it is preferentially determined as the outflow of the power generation tail water, and the remaining water volume is determined as the outflow of the flood discharge facilities; otherwise, all are determined as the outflow of the flood discharge facilities.

[0074] In the present invention, an implementation of the mathematical model of the open - discharge control operation mode of the spillway gate of the reservoir project is as follows: (1) Select the discharge capacity curve. If the initial water level at the beginning of the scheduling time period exceeds the maximum controlled - discharge water level in the project scheduling regulations, select the maximum sum of the flow - through capacities of all spillways (open - discharge line); otherwise, select the maximum flow - through capacity curve of the common spillway. (2) When the initial water level of the time period is lower than the crest elevation of the spillway / bottom elevation of the gate, give priority to water storage. (3) When the initial water level of the time period is greater than or equal to the crest elevation of the spillway / bottom elevation of the gate, considering the relationship between the incoming water in the current time period and the maximum discharge capacity corresponding to the initial water level, if the incoming water is large, iteratively calculate a higher water level at the end of the time period based on the principle of water balance and the dichotomy method, and vice versa, obtain a lower water level at the end of the time period. (4) For the calculated time period discharge, if the initial and end water levels of the time period both meet the constraints of the power generation diversion water level range, it is preferentially determined as the outflow of the power generation tail water, and the remaining water volume is determined as the outflow of the flood discharge facilities; otherwise, all are determined as the outflow of the flood discharge facilities.

[0075] In the present invention, an implementation of the mathematical model of the upstream water level control operation mode of the sluice dam project is as follows: (1) Perform time interpolation according to the specified target water level control parameters to obtain the target water level at the end of the current time period. Based on the water level in front of the sluice at the beginning of the time period of the sluice, judge the highest and lowest water levels in front of the sluice at the end of the time period through the minimum and maximum discharge capacities, and obtain the reasonable range of the water level in front of the sluice at the end of the time period. Subsequently, select the value closest to the target water level at the end of the time period within the reasonable range as the water level at the end of the time period. (2) When the initial water level is less than or equal to the flood limit water level + super - elevation, give priority to water storage to obtain the upper limit of the water level in front of the sluice at the end of the time period; the sluice iteratively calculates the lower limit of the water level at the end of the time period with the maximum capacity for discharge. (3) When the initial water level is greater than the flood limit water level + super - elevation, the sluice iteratively calculates the water level at the end of the time period with the maximum capacity for discharge, and this water level serves as both the upper limit and the lower limit. (4)For the calculated hourly discharge, if the water levels at the beginning and end of the hour both meet the constraints of the power generation diversion water level range, it is preferentially determined as the outflow of the power generation tail water, and the remaining water volume is determined as the outflow of the flood discharge facilities; otherwise, all are determined as the outflow of the flood discharge facilities.

[0076] In the present invention, an implementation of the mathematical model for the gradually open flood control operation mode of the sluice dam project is as follows: (1)If any of the following conditions is satisfied, it is determined that the sluice dam at the current time step is in a fully open state: 1) The inflow in the hour is greater than the maximum discharge capacity corresponding to the water level at the beginning of the hour; 2) The water level at the beginning of the hour is greater than the flood limit water level + safety margin; 3) The sluice dam in the previous hour was in an open state. (2)If the current hour is in a fully open state, considering the relationship between the inflow in the current hour and the maximum discharge capacity corresponding to the initial water level, if the inflow is large, the higher water level at the end of the hour is iteratively calculated based on the principle of water balance and the dichotomy method; otherwise, the lower water level at the end of the hour is obtained. (3)If the current hour is still in a gradually open state, the lower water level at the end of the hour is iteratively calculated. If the water level change range in the hour is less than or equal to the maximum allowable change range in the dispatching rules, the sluice dam at the end of the hour is adjusted to a fully open state; otherwise, the water level at the end of the hour is corrected as follows: the initial water level minus the maximum allowable change range. (4)For the calculated hourly discharge, if the water levels at the beginning and end of the hour both meet the constraints of the power generation diversion water level range, it is preferentially determined as the outflow of the power generation tail water, and the remaining water volume is determined as the outflow of the flood discharge facilities; otherwise, all are determined as the outflow of the flood discharge facilities.

[0077] VIII. Industry Application Value Topological analysis: Traditional methods rely on manual drawing analysis, which is time-consuming and difficult to adapt to engineering dispatching changes. Through matrix automatic construction and algorithm division, the present invention completes topological analysis within 10 seconds and can dynamically adapt to dispatching changes.

[0078] Control simulation: Traditional single-project serial simulation takes 2 - 3 times as long for a single project and has a single mode. The present invention performs parallel computing according to the dominant camps, supports multiple control modes, and the overall efficiency is increased by more than 3 times, and the simulation flexibility is greatly improved.

[0079] Optimization iteration: Traditional methods rely on empirical trial calculations, and more than 2 hours are required for 100 iterations, and it is easy to fall into local optima. The present invention combines intelligent evolution with an acceleration strategy, reducing the number of iterations by 40%, demonstrating the advantages of the engineering regulation intelligent optimizer and the calculation acceleration method.

[0080] This embodiment details the operation process of each component and the application of the calculation acceleration method, and verifies the effectiveness of the system through actual projects. Each component works closely together, and the calculation acceleration method significantly improves the calculation efficiency, providing strong support for the scientific regulation of the basin water project group, and fully demonstrating the innovation and practicality of the present invention in the field of water conservancy projects.

[0081] The present invention realizes the intelligent construction of the partition set and dominance relationship of basin water projects by introducing a basin topology intelligent processor, laying a foundation for parallel computing; combines a fine simulator for reservoir and sluice control plans, a scheduling plan simulator, and a flow simulator for flood storage and detention spaces to construct a multi-level and high-precision simulation system for water projects and water flow processes; at the same time, the collaborative work of the multi-objective evaluator for scheduling plans and the intelligent optimizer for project regulation realizes the quantitative evaluation and intelligent optimization of regulation plans. In addition, the present invention innovatively proposes a calculation acceleration method, which significantly improves the system operation efficiency through parallel computing, efficient invocation of the historical plan library, and avoidance of invalid calculations.

[0082] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. An optimized regulation system for the control and operation of basin water projects for intelligent water conservancy, characterized in that: It includes a basin topology intelligent processor, a fine simulator for reservoir and sluice control plans, a simulator for reservoir and sluice operation plans, a flow simulator for flood storage and detention spaces, a multi-objective evaluator for operation plans, and an intelligent optimizer for project regulation and control; The basin topology intelligent processor is used to intelligently construct the neighborhood domination matrix D1 and the global domination matrix D2 of the basin reservoir and sluice project group, analyze and obtain the upstream control and operation project set S1, the optimized regulation project set S2, the downstream control and operation project set S3, and the domination camps within the sets, to support the parallel computing of the simulation analysis of the basin project group; it is used to obtain the specified operation modes and parameters of each reservoir and sluice project in the basin, and obtain the inflow of each project and its evolution lag time statement; The fine simulator for reservoir and sluice control plans calculates the synthetic inflow process of the project according to the specified inflow of the control and operation project and its evolution lag time statement; The simulator for reservoir and sluice operation plans calculates the synthetic inflow process of the project according to the specified inflow of the optimized regulation project and its evolution lag time statement; The flow simulator for flood storage and detention spaces identifies the most downstream project in the basin reservoir and sluice project that does not dominate other projects through the global domination matrix D2. Facing the downstream flood storage and detention space of the most downstream project, with the outflow calculated by the most downstream project as the upstream incoming water boundary, based on the river channel confluence hydrology method or the hydrodynamic method, it conducts continuous time-step simulation of the flow in the flood storage and detention space, outputs the hydraulic element process of the control section in the flood storage and detention space within the calculation time domain, and quantitatively determines the constraint violation amount of the flood storage and detention space corresponding to the current upstream incoming water boundary. The involved constraint conditions include: hydraulic element range constraint, hydraulic element amplitude change constraint, hydraulic element rate of change constraint, hydraulic element threshold damage depth constraint, and hydraulic element threshold damage duration constraint; The multi-objective evaluator for operation plans conducts multi-objective evaluation on the current optimized regulation project operation plan; The intelligent optimizer for project regulation and control consists of three parts: a plan population initialization unit, a plan population quantitative evaluation unit, and a plan population heuristic evolution unit; among them, the plan population initialization unit randomly generates a two-dimensional matrix. Each row of the matrix represents a scheduling plan for all projects participating in the optimized regulation, that is, the upstream water level or the downstream discharge process of each project within the calculation time domain. The matrix is formed by combining multiple scheduling plans in columns, which is the plan population; the plan population quantitative evaluation unit assigns a quantitative priority to each scheduling plan using an adaptive constraint handling method based on the weighted objective deviation amount and the total constraint violation amount of each scheduling plan in the population; the plan population heuristic evolution unit iteratively updates and evolves the plan population using a heuristic evolution algorithm based on the quantitative priorities of each scheduling plan.

2. The optimized regulation system for the control and operation of basin water projects for intelligent water conservancy according to claim 1, characterized in that: The described basin topology intelligent processor constructs a neighborhood domination matrix D1 of the basin project, where D1 is a two-dimensional square matrix with dimensions of M + N. M is the number of reservoir projects, and N is the number of sluice projects. The value of D1ij being 1 indicates that the outflow of project i directly flows into the adjacent downstream project j, that is, project i affects and dominates project j. The value of -1 indicates that project i directly receives the outflow of the adjacent upstream project j, that is, project i is affected and dominated by project j. The value of 0 indicates that there is no mutual influence and domination relationship between project i and project j. Based on the neighborhood domination matrix D1, matrix analysis is carried out to obtain the global domination matrix D2. D2 is a two-dimensional square matrix with dimensions of M + N. The value of D2ij being 1 indicates that the outflow of project i directly or indirectly flows into the downstream project j, that is, project i affects and dominates project j. The value of -1 indicates that project i directly or indirectly receives the outflow of the upstream project j, that is, project i is affected and dominated by project j. The value of 0 indicates that there is no mutual influence and domination relationship between project i and project j.

3. A regulation method for the optimized regulation system for the control and operation of basin water projects for intelligent water conservancy as claimed in claim 1 or 2, characterized in that, It includes the following steps: Step 1: Apply the described watershed topology intelligent processor to obtain the specified scheduling methods and parameters of each reservoir and sluice project in the watershed, the inflow of each project and its evolution lag time statement, construct the watershed project neighborhood domination matrix D1 and the global domination matrix D2, divide the watershed reservoir and sluice group into S1, S2, and S3. S1 is the set of upstream control operation projects, S2 is the set of optimized regulation projects, and S3 is the set of downstream control operation projects. Use the fast non-dominated sorting algorithm to obtain multiple domination camps within each project set. The outflow of the projects contained in the upper camp is used as the inflow of the projects contained in the lower camp. For all projects belonging to the same camp, there is no intersection between their outflows and inflows. Record the CPU time at the start of Step 1 ; Step 2: Based on the above-mentioned reservoir and sluice control plan fine simulator, for the upstream control operation project set S1, project simulation calculations are carried out for each domination camp one by one: First, calculate the current project synthetic inflow process according to the inflow and its evolution lag time statement. Then, apply the mathematical models of various control operation modes of different types of projects, and combine with the current project control parameters to carry out continuous calculation time step simulation, and output the upstream water level of the current project and the calculated outflow process within the calculation time domain. Step 3: Based on the above-mentioned project regulation intelligent optimizer, use the plan population initialization unit to randomly generate a plan population, that is, a two-dimensional matrix where each row represents a scheduling plan for all projects participating in the optimization regulation project. The scheduling plan is the upstream water level or the downstream discharge process of the project within the calculation time domain. Step 4: Construct an empty historical plan library to store the evaluated optimization regulation project scheduling plans, as well as the corresponding reservoir and sluice project and flood storage and detention space simulation results, weighted target deviation amount, and total constraint violation amount. Step 5: For each scheduling plan of the optimization regulation project in the current plan population, if an evaluated scheduling plan can be located in the historical plan library, directly retrieve the reservoir and sluice project and flood storage and detention space simulation results, weighted target deviation amount, and total constraint violation amount; otherwise, carry out simulation calculations to obtain the reservoir and sluice project and flood storage and detention space simulation results, weighted target deviation amount, and total constraint violation amount, and add them to the historical plan library. Step 6: Use the plan population quantitative evaluation unit of the above-mentioned project regulation intelligent optimizer to assign a quantitative priority to each scheduling plan based on the weighted target deviation amount and total constraint violation amount of each scheduling plan in the population by using an adaptive constraint processing method. The specific calculation method is as follows: ； Among them, is the priority of a scheduling plan in the population, is the weighted objective deviation of this scheduling plan, is the total constraint violation amount of this scheduling plan, is the proportion of feasible plans with a total constraint violation amount of 0 in the population, is the deviation of each scheduling objective of the optimized regulation project and the flood storage and detention space is the weight coefficient, is the violation amount of each constraint of the optimized regulation project and the flood storage and detention space; Step 7: Use the pre - plan population heuristic evolution unit of the engineering control intelligent optimizer to update and optimize the engineering scheduling pre - plan population using the heuristic evolution algorithm based on the quantitative priorities of each scheduling pre - plan, and read the CPU time at this moment ; Repeat steps 5 to 7 until the exit condition is met. The exit conditions include that the total calculation time reaches the preset upper limit or the number of iterative updates reaches the preset upper limit. Calculate the total calculation time : ; The number of iterative updates is calculated as follows: the number of times steps 5 to 7 are repeatedly executed; after exiting, output the optimization and regulation engineering scheduling plan with the highest priority during the iterative calculation process.

4. The regulation method for the optimized regulation system for the control and operation of basin water projects for intelligent water conservancy according to claim 3, characterized in that: The method of applying the fast non-dominated sorting algorithm to the domination matrices D1 and D2 is as follows: A11. If no project in the basin participates in the optimization regulation project, all projects belong to the upstream control operation project set S1; if at least one project in the basin participates in the optimization regulation project, the optimization regulation project set S2 is not empty. If a control operation project appears within the scope of two optimization and regulation projects, the set of participating optimization and regulation projects is a subset of the optimization and regulation project set S2; A12. If the optimization and regulation project set S2 is not empty, starting from all the participating optimization and regulation projects, continuously locate the adjacent downstream projects according to matrices D1 and D2 and store them in the union of S2 and S3. After obtaining the union of S2 and S3, use the element exclusion method to obtain the upstream control operation project set S1; A13. According to matrices D1 and D2, determine the reservoir and dam projects that do not dominate other projects at all, do not belong to the upstream control operation project set S1, and are designated to carry out control operations. Starting from these reservoir and dam projects, continuously locate the adjacent upstream projects and store them in the downstream control operation project set S3. The condition that the adjacent upstream projects should meet is that they do not belong to the upstream control operation project set S1 and are designated to carry out control operations. If there are no adjacent upstream projects that meet the conditions during the location process, the location ends; A14. On the basis of obtaining the upstream control operation project set S1 and the downstream control operation project set S3, use the element exclusion method to obtain the optimization and regulation project set S2; A15. For the sets S1, S2, and S3 respectively, extract the corresponding rows and columns of matrices D1 and D2 as new temporary dominance matrices, and apply the fast non-dominated sorting algorithm to divide each project set into different dominance camps.

5. The regulation method for the optimized regulation system for the control and operation of basin water projects for intelligent water conservancy according to claim 4, characterized in that: In step 5 mentioned above, the simulation calculation includes the following steps: Step 5.1: For each dominance camp in the optimization and regulation project set S2, carry out project type judgment and simulation calculation in the following way: If the current project is a control operation project, based on the refined simulator of the reservoir and dam control plan, calculate the synthetic inflow process of the current project according to the inflow and its evolution lag declaration, apply the mathematical models of various control operation modes of different types of projects, and perform continuous calculation time step simulation in combination with the control parameters of the current project, and output the upstream water level and calculated outflow process of the current project within the calculation time domain; If the current project is an optimization and regulation project, based on the reservoir and dam operation plan simulator, calculate the synthetic inflow process of the current project according to the inflow and its evolution lag declaration, based on the operation plan of the current project, perform continuous calculation time step simulation in combination with the optimization parameters, and output the upstream water level, calculated inflow, and calculated outflow process of the current project within the calculation time domain, and quantify the constraint violation amount of each project at each calculation time step; Step 5.2: Based on the refined simulator of the reservoir and dam control plan, for the downstream control operation project set S3, carry out simulation calculation for each project in each dominance camp: First, calculate the synthetic inflow process of the current project according to the inflow and its evolution lag declaration, and then apply the mathematical models of various control operation modes of different types of projects, and perform continuous calculation time step simulation in combination with the control parameters of the current project, and output the upstream water level and calculated outflow process of the current project within the calculation time domain; Step 5.3: If the constraint violation amounts of all projects at each calculation time step in Step 5.1 are all 0, based on the above-mentioned flood detention space flow simulator, with the calculated outflow process of the project at the most downstream of the basin as the upstream incoming water boundary, conduct continuous calculation time step simulation of the flood detention space flow, output the hydraulic element process of the control section of the flood detention space within the calculation time domain, and quantitatively determine the constraint violation amount of the flood detention space; otherwise, directly assign a very large value to the constraint violation amount of the flood detention space; Step 5.4: Based on the above-mentioned multi-objective evaluator of the operation plan, conduct multi-objective evaluation for the current optimized regulation project operation plan. On the basis of quantitatively determining the constraint violation amount of the optimized regulation project and the constraint violation amount of the flood detention space, quantitatively determine the deviation of each scheduling objective of the optimized regulation project and the flood detention space, and then comprehensively determine the weighted objective deviation amount and the total constraint violation amount; Step 5.5: Add the simulation results of the reservoir sluice projects and the flood detention space, the weighted objective deviation amount and the total constraint violation amount of the current simulated evaluation operation plan to the historical plan library.

Images