Cascade reservoir flood control capacity step-by-step release optimization scheduling method and system
The optimization scheduling problem of flood control capacity release of cascade reservoirs is decomposed through the gradient-like gradual optimization method, which solves the problems of local convergence and dimensional disasters of the existing methods, and achieves efficient and accurate optimization scheduling of flood control capacity release of reservoirs.
Patent Information
- Application Number
- CN202411005563.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-07-25
AI Technical Summary
The existing cascade reservoir flood control capacity release optimization scheduling method has problems such as local convergence and dimensional disasters, which is difficult to meet the requirements of computing efficiency and solution accuracy.
By using the gradual optimization method of class gradient, by constructing objective functions and constraints, combining manual empirical decision-making and class gradient depth search, we decompose multi-stage problems and perform class gradient depth search at each stage, generating diverse search information to improve local search ability.
The rapid and precise solution of the gradual release of optimized scheduling of the flood control capacity of the cascade reservoir is achieved, which avoids dimensional disaster problems and improves the local deep search accuracy and calculation efficiency.
Smart Images

Figure CN119047732A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of reservoir group optimization dispatching, and in particular to a method and system for optimizing dispatching of cascade reservoirs for flood control by gradually releasing storage capacity. Background Art
[0002] In order to promote benefits and eliminate harm, store flood water to compensate for drought, and regulate runoff, reservoir projects have been built and put into operation one after another, and through the implementation of scientific and precise dispatching of reservoir groups, they have provided important support for protecting the lives and property of the people and promoting the stable and healthy development of the economy and society. However, at present, most reservoirs share the storage capacity for flood control and benefit development, and one reservoir is used for multiple development tasks such as flood control, power generation, water supply, shipping, and ecology, making it difficult to coordinate the joint flood control and water storage and release of reservoir groups. It is necessary to coordinate the relationship between flood control and benefit development to promote the full realization of the comprehensive benefits of reservoir groups.
[0003] As the scale of reservoir groups continues to expand, the ability of reservoir groups to regulate floods is gradually increasing. In addition, the level of hydrological forecasting technology is constantly improving, and the laws of flood occurrence are gradually becoming clear. In order to promote the effective use of flood resources, it is very important to implement the gradual release of flood control storage capacity of reservoir groups on the premise of ensuring flood control safety. In other words, combined with the favorable situation that the overall flood control capacity of the basin has been greatly improved after the use of large-scale reservoir groups, on the premise of ensuring flood control safety in the basin, it is possible to propose an optimization control plan for the operating water level of cascade reservoirs during the flood season according to the flood season water situation, thereby providing conditions for the effective use of flood resources in cascade reservoirs during the flood season, improving the water storage rate at the end of the flood season, and even giving full play to the comprehensive benefits of cascade reservoirs. To this end, it is of great practical significance to optimize the gradual release process of flood control storage capacity of cascade reservoirs during the flood season in combination with the law of flood changes during the flood season.
[0004] In essence, the process of gradually releasing the flood control storage capacity of cascade reservoirs is also to control the operating water level of cascade reservoirs during the flood season, which is a key means to coordinate flood control and economic development. In order to achieve the optimization of economic development benefits while ensuring flood control safety, it is necessary to carry out research on the optimal scheduling of the gradual release of flood control storage capacity of cascade reservoirs. From a mathematical point of view, the optimal scheduling of the gradual release of flood control storage capacity of cascade reservoirs is similar to the optimal scheduling of power generation and flood control of cascade reservoirs, and also has the characteristics of multi-stage, high dimension, strong constraints, and nonlinearity.
[0005] As an improved method of traditional dynamic programming method, the step-by-step optimization method (POA) decomposes the multi-stage optimization problem into several two-stage sub-problems for calculation, and then iterates and optimizes only the current stage in each calculation, and calculates stage by stage until the final solution is output. POA has been widely used in many fields such as hydropower scheduling and flood control scheduling, but it is also found in applications that as the scale of the system increases, POA still has problems such as poor search ability and long calculation time, which makes it difficult to meet the requirements of accuracy and effectiveness in solving optimization problems.
[0006] The traditional gradient method is a common method for solving nonlinear optimization problems. It gradually approaches the optimal solution along the effective direction, but the prerequisite is that the objective function is differentiable within the solution range. However, since most of the characteristic curves used in the optimal operation of a reservoir group are measured point sets, the constraints and objective functions are nonlinear and difficult to be directly derived. To this end, based on the gradient method, related studies have proposed a discrete gradient descent method for the optimal operation of a reservoir group (Zhao Zhipeng, Liao Shengli, Cheng Chuntian, et al. Discrete gradient stepwise optimization algorithm for the medium- and long-term optimal operation of cascade hydropower stations [J]. Journal of Hydraulic Engineering, 2018, 49(10): 1243-1253.). However, each iterative calculation of this method requires the calculation of discrete gradient values, which makes the calculation process very complicated. In addition, the search step size is not easy to determine. Whether the step size is too large or too small, it will affect the calculation accuracy.
[0007] Therefore, the applicant considered proposing a gradient-like stepwise optimization method to solve the problem of local convergence of POA when facing the optimal scheduling problem of cascade reservoir flood control storage capacity release. Summary of the invention
[0008] In order to overcome the shortcomings of the above-mentioned technology, the present invention provides a method and system for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs, which solves the problems of local convergence and dimensionality curse faced by POA when solving the optimization scheduling problem of the release of flood control storage capacity of cascade reservoirs, and has the advantages of high computational efficiency and high solution accuracy.
[0009] To achieve the above purpose, the technical solution adopted by the present invention is as follows:
[0010] A method for optimizing the scheduling of gradually releasing the flood control storage capacity of cascade reservoirs comprises the following steps:
[0011] 1) Collect basic information on flood season operation of cascade reservoirs and joint operation plan of cascade reservoirs;
[0012] 2) According to the joint dispatching scheme of cascade reservoirs, obtain the releasable flood control storage capacity of cascade reservoirs at each stage as the flood characteristics change; based on the basic data and the releasable flood control storage capacity, construct the objective function and constraint conditions of the gradual release of the flood control storage capacity of cascade reservoirs, and set the calculation parameters; the calculation parameters include the number of reservoirs N, the number of dispatching stages T, the maximum number of iterations M, and the maximum number of class gradient searches K;
[0013] 3) Using artificial experience decision-making or conventional methods to obtain the initial state of the water level process of gradually releasing the flood control storage capacity of the cascade reservoirs that meets the constraints And calculate the discrete step length of each reservoir at each stage Δ=(Δ i,j ) N×T , the calculation formula is as follows:
[0014]
[0015] In the formula, and Δ i,j They represent the initial state of reservoir i at stage j respectively; and Z i,j Respectively represent the upper and lower limits of the water level of reservoir i at stage j; i = 1, 2, ..., N; j = 1, 2, ..., T;
[0016] 4) Initialize the number of iterations m = 1;
[0017] 5) performing a gradient-like stepwise optimization method;
[0018] 6) Let m=m+1, if m>M, go to step 7); otherwise go to step 5) and continue to perform gradient-like depth search;
[0019] 7) Stop the calculation and output the final optimal trajectory state, that is, obtain the optimal dispatching process of gradually releasing the flood control storage capacity of the cascade reservoirs with the best benefits.
[0020] Preferably, in step 1), the basic data include characteristic parameters of each reservoir, water level storage capacity curve, downstream water level flow relationship curve, power station unit output characteristic curve, and reservoir water inflow process.
[0021] Preferably, in step 2), the objective function is:
[0022]
[0023] Where: F is the total power generation during the gradual release of the flood control storage capacity of the cascade reservoirs; N is the number of reservoirs; i is the reservoir number, and i = 1, 2, ..., N; T is the number of stages in the scheduling period during which the flood control storage capacity of the cascade reservoirs is gradually released; j is the stage number, and j = 1, 2, ..., T; A i is the output coefficient of reservoir i; Q i,j is the power generation flow of reservoir i in stage j, m 3 / s;H i,j is the average power generation head of reservoir i in stage j after deducting the head loss, m; Δt is the stage duration, h.
[0024] Preferably, in step 2), the constraints include:
[0025] (1) Water balance constraints of each reservoir: V i,j =V i,j-1 +(I i,j-1 -O i,j-1 )×Δt,
[0026] (2) Hydraulic connection constraints of each reservoir: I i,j =Oi-1,j +R i,j ,
[0027] (3) Water level constraints of each reservoir:
[0028] (4) Constraints on outflow from each reservoir: O i,j =Q i,j +S i,j ,
[0029] (5) Constraints on discharge capacity of each reservoir: i,j ≤O i,max (Z i,j ),
[0030] (6) Output constraints of each reservoir: P min i,j ≤A i Q i,j ·H i,j ≤P max i,j ,
[0031] (7) Cascade reservoirs can release flood control storage capacity constraints:
[0032] (8) Constraints on total output of cascade reservoirs:
[0033] (9) Non-negative constraints: All variables are non-negative;
[0034] Where V i,j is the storage capacity of reservoir i at the end of stage j, m 3 ;I i,j is the inflow of reservoir i in stage j, m 3 / s;O i,j is the outflow of reservoir i in stage j, m 3 / s; and are the minimum outflow and maximum outflow of reservoir i in stage j, respectively, 3 / s;S i,j is the water discharge of reservoir i at stage j, m 3 / s; R i,j is the interval flow of reservoir i in stage j, m 3 / s; Z i,j is the water level before the dam of reservoir i at stage j, m; and are the lowest and highest water levels in front of the dam of reservoir i at stage j, m; O i,max (Z i,j ) is the water level Z corresponding to reservoir i at stage j i,j Maximum discharge capacity, m 3 / s;P min i,j and P max i,j are the minimum and maximum output of reservoir i at stage j, kW; V i 0 is the storage capacity corresponding to the flood control limit water level of reservoir i, m 3 SV j is the total releasable flood control storage capacity of the cascade reservoirs at stage j, m 3 NP j is the minimum total output of the cascade reservoirs at stage j, kW.
[0035] Preferably, the step 5) comprises the following steps:
[0036] 5.1) Decompose the optimization scheduling problem of gradually releasing flood control reservoir capacity into T-1 two-stage problems;
[0037] 5.2) For each two-stage problem, perform gradient-like deep search in turn.
[0038] Preferably, the step 5.2) comprises the following steps:
[0039] 5.2.1) Initialize the number of class gradient searches k = 1;
[0040] 5.2.2) Status Perform gradient-like deep search to generate gradient-like states in, is the state of the sth two-stage at the mth iteration;
[0041] 5.2.3) Comparison of class gradient states and current status if Better than The current state is replaced by the gradient-like state, otherwise no processing is done;
[0042] 5.2.4) Let k = k + 1. If k ≤ K, it indicates that the gradient-like depth search has been completed in the current stage, and go to step 5.2.5). Otherwise, go to step 5.2.2) and continue to perform the gradient-like depth search.
[0043] 5.2.5) Determine whether all two-stage sub-problems have executed gradient-like deep search. If so, go to step 6), otherwise, go to step 5.2.1) and execute gradient-like deep search on the remaining two-stage sub-problems in turn.
[0044] Preferably, the gradient-like state The calculation formula is as follows:
[0045]
[0046] In the formula, α1, α2, ..., α N-1 are N-1 random angles with values in [0,2π]. According to the characteristics of trigonometric functions, the gradient-like equidistant spatial parameters D1, D2, ..., D N The value of is between -1 and 1, ensuring At the same time, g is the step size control parameter, which is set to gradually decrease with the advancement of iteration to gradually shrink the search range of the class gradient.
[0047] Preferably, the calculation formula of the step length control parameter g is: g=1-m / M.
[0048] A cascade reservoir flood control storage capacity gradual release optimization dispatching system, comprising:
[0049] An initialization module is used to construct an optimization objective function and constraint conditions for the gradual release of the flood control storage capacity of cascade reservoirs, and generate a water level trajectory for the gradual release of the flood control storage capacity of cascade reservoirs that meets the constraint conditions based on artificial experience decision-making or conventional dynamic programming methods;
[0050] The stepwise optimization decomposition module is used to decompose the optimization scheduling problem of the stepwise release of the flood control storage capacity of the cascade reservoirs at the current iteration number into multiple two-stage stepwise optimization sub-problems;
[0051] The gradient-like deep search module performs gradient-like deep search based on the step-by-step optimization decomposition module, including gradient-like parameter generation, gradient-like state generation, state comparison and replacement, and stop condition judgment functions to obtain a better optimization state;
[0052] The output module is used to repeatedly execute the step-by-step optimization decomposition module and the gradient-like deep search module until the preset iteration stop condition is met to obtain the optimal scheduling result, which serves as the water level optimization process for gradually releasing the flood control storage capacity of the final cascade reservoirs.
[0053] A computer device comprises a memory and a processor, wherein the memory is used to store at least one program, and the processor is used to load the at least one program to execute the above-mentioned method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs.
[0054] Compared with the prior art, the present invention has the following beneficial effects:
[0055] The method of the present invention has clear principles, simple calculations, and is easy to implement, and can achieve rapid and accurate solutions to the optimization scheduling problem of gradually releasing the flood control storage capacity of cascade reservoirs. The present invention innovatively introduces a quasi-gradient method into the solution process of each stage of POA, and combines equidistant spatial parameters based on angle patterns and step control parameters based on iterative processes to generate diverse search information, so as to improve the local search capability of POA and seek the optimization direction of the scheduling trajectory. The method of the present invention performs optimization according to the quasi-gradient method in the optimization process of each stage, avoids the comprehensive combination of all discrete states of each reservoir at each stage, avoids the problem of the curse of dimensionality, and improves the accuracy of local deep search. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] Figure 1 It is a flow chart of a method for optimizing the scheduling of the flood control storage capacity of cascade reservoirs gradually released according to the present invention;
[0057] Figure 2 It is a comparison diagram of the power generation change process of the method of the present invention and the POA method in the optimized scheduling of the flood control storage capacity of the cascade reservoirs in a normal water year;
[0058] Figure 3 It is a water level process diagram of reservoir A in the optimized scheduling of the gradual release of flood control storage capacity of the cascade reservoirs in a normal water year in an embodiment of the present invention;
[0059] Figure 4 It is a water level process diagram of Reservoir B in the optimized scheduling of the gradual release of flood control storage capacity of the cascade reservoirs in a normal water year in an embodiment of the present invention;
[0060] Figure 5 It is a water level process diagram of reservoir C in the optimized scheduling of the gradual release of the flood control storage capacity of the cascade reservoirs in a normal water year in an embodiment of the present invention;
[0061] Figure 6 It is a diagram of the water level process of the D reservoir in the optimized scheduling of the gradual release of the flood control storage capacity of the cascade reservoirs in a normal water year in an embodiment of the present invention. DETAILED DESCRIPTION
[0062] In order to better explain the present invention, the main contents of the present invention are further explained below in conjunction with the drawings and specific embodiments, but the contents of the present invention are not limited to the following embodiments.
[0063] like Figure 1 As shown, the present invention provides a method for optimizing the scheduling of gradually releasing the flood control storage capacity of cascade reservoirs, comprising the following steps:
[0064] 1) Collect basic data on flood season dispatching of cascade reservoirs, including characteristic parameters of each reservoir, water level and storage capacity curve, downstream water level and flow relationship curve, hydropower station unit output characteristic curve, reservoir water inflow process and other data and dispatching documents such as cascade reservoir joint dispatching plan.
[0065] 2) According to the joint dispatching plan of cascade reservoirs, the releasable flood control storage capacity of the cascade reservoirs at each stage that changes with flood characteristics is obtained; based on the basic data and the releasable flood control storage capacity, the objective function and constraint conditions of the optimal dispatching of the gradual release of the flood control storage capacity of the cascade reservoirs are constructed, and the calculation parameters are set, including the number of reservoirs N, the number of dispatching stages T, the maximum number of iterations M, and the maximum number of class gradient searches K.
[0066] According to the cascade reservoir scheduling plan, on the premise that the cascade reservoirs can release flood control storage capacity at different periods of the flood season, in order to coordinate the release rhythm of the flood control storage capacity of each reservoir, improve the benefits of the cascade reservoirs, and improve the efficiency of flood resource utilization during the flood season, the constructed cascade reservoir flood control storage capacity gradual release optimization scheduling model mainly considers: the inflow process of each reservoir at different scheduling stages of the flood season is known, and on the premise of meeting the flood control safety of the basin and the requirements of the flood control storage capacity that can be released by the cascade reservoirs at each stage, the flood control storage capacity of the cascade reservoirs is gradually released, thereby achieving maximum power generation.
[0067] To this end, the objective function of the optimization scheduling model for the gradual release of flood control storage capacity of cascade reservoirs is as follows:
[0068]
[0069] Where: F is the total power generation during the gradual release of the flood control storage capacity of the cascade reservoirs; N is the number of reservoirs; i is the reservoir number, and i = 1, 2, ..., N; T is the total number of stages in which the flood control storage capacity of the cascade reservoirs is gradually released; j is the stage number, and j = 1, 2, ..., T; A i is the output coefficient of reservoir i; Q i,j is the power generation flow of reservoir i in stage j (m 3 / s); H i,j is the average power generation head of reservoir i in stage j after deducting the head loss (m); Δt is the stage duration (h).
[0070] The various constraints that need to be met for the gradual release of flood control storage capacity of cascade reservoirs for optimal scheduling mainly include:
[0071] (1) Water balance constraints of each reservoir: V i,j =V i,j-1 +(I i,j-1 -O i,j-1 )×Δt,
[0072] (2) Hydraulic connection constraints of each reservoir: Ii,j =O i-1,j +R i,j ,
[0073] (3) Water level constraints of each reservoir:
[0074] (4) Constraints on outflow from each reservoir: O i,j =Q i,j +S i,j ,
[0075] (5) Constraints on discharge capacity of each reservoir: i,j ≤O i,max (Z i,j ),
[0076] (6) Output constraints of each reservoir: P min i,j ≤A i Q i,j ·H i,j ≤P max i,j ,
[0077] (7) Cascade reservoirs can release flood control storage capacity constraints:
[0078] (8) Constraints on total output of cascade reservoirs:
[0079] (9) Non-negative constraint: All variables are non-negative values.
[0080] Where V i,j is the storage capacity of reservoir i at the end of stage j (m 3 );I i,j is the inflow of reservoir i at stage j (m 3 / s); i,j is the outflow of reservoir i at stage j (m 3 / s); and are the minimum outflow of reservoir i in stage j (m 3 / s) and maximum outbound flow (m 3 / s); S i,j is the water discharge of reservoir i at stage j (m 3 / s); R i,j is the interval flow of reservoir i in stage j (m 3 / s); Z i,j is the water level before the dam of reservoir i at stage j (m); and are the lowest and highest water levels in front of the dam of reservoir i at stage j (m); O i,max (Z i,j ) is the water level Z corresponding to reservoir i at stage j i,j Maximum discharge capacity (m 3 / s); P min i,j and P max i,j are the minimum output (kW) and maximum output (kW) of reservoir i at stage j; V i 0 is the storage capacity corresponding to the flood control limit water level of reservoir i (m 3 );SV j is the total releasable flood control storage capacity of the cascade reservoirs at stage j (m 3 );NP j is the minimum total output of the cascade reservoirs in stage j (kW).
[0081] 3) Using artificial experience decision-making or conventional methods to generate the initial state of the water level process of the gradual release of the flood control storage capacity of the cascade reservoirs that meets various constraints And calculate the discrete step length of each reservoir at each stage Δ=(Δ i,j ) N×T , the calculation formula is as follows:
[0082]
[0083] In the formula, and Δ i,j They represent the initial state of reservoir i at stage j respectively; and Z i,j They represent the upper and lower limits of the water level of reservoir i at stage j respectively.
[0084] 4) Initialize the number of iterations m = 1,
[0085] 5) Execute the gradient-like stepwise optimization method, the calculation steps of which are as follows:
[0086] 5.1) Decompose the optimization scheduling problem of gradually releasing flood control reservoir capacity into T-1 two-stage problems;
[0087] 5.2) For each two-stage problem, perform gradient-like deep search in turn.
[0088] The basic principle of gradient-like deep search is to gradually approach the optimal solution along the effective direction based on the gradient method. It uses individuals to search along several equidistant spaces and randomly generates equidistant spaces based on angle patterns to obtain the optimal evolutionary direction until the iterative search is completed.
[0089] Now take the state of the sth two-stage in the mth iteration Taking the gradient-like deep search optimization problem as an example, we will explain it in detail.
[0090] 5.2.1) Initialize the number of class gradient searches k = 1;
[0091] 5.2.2) Status Perform gradient-like deep search to generate gradient-like states The calculation formula is as follows:
[0092]
[0093] In the formula, α1, α2, ..., α N-1 are N-1 random angles with values in [0,2π]. According to the characteristics of trigonometric functions, the gradient-like equidistant spatial parameters D1, D2, ..., D N The value of is between -1 and 1, ensuring The up and down fluctuations are traversed to find the best solution; at the same time, g is the step size control parameter, which is set to gradually decrease with the advancement of iteration to gradually shrink the search range of the class gradient. The calculation formula is as follows:
[0094] g = 1-m / M;
[0095] 5.2.3) Comparison of class gradient states and current status if Better than The current state is replaced by the gradient-like state; otherwise, no processing is performed.
[0096] 5.2.4) Let k = k + 1. If k ≤ K, it indicates that the class gradient depth search has been completed in the current stage, and go to step 5.2.5); otherwise, go to step 5.2.2) and continue to perform the class gradient depth search.
[0097] 5.2.5) Determine whether all two-stage sub-problems have executed gradient-like depth search. If so, go to step 6); otherwise, go to step 5.2.1) and execute gradient-like depth search on the remaining two-stage sub-problems in turn.
[0098] 6) Let m=m+1. If m>M, go to step 7); otherwise, go to step 5) and continue to perform gradient-like depth search.
[0099] 7) Stop the calculation, output the final optimal trajectory state, and obtain the optimal dispatching process of gradually releasing the flood control storage capacity of the cascade reservoirs that takes into account the optimal benefits, that is, obtain the optimal dispatching process of gradually releasing the flood control storage capacity of the cascade reservoirs with the largest total power generation.
[0100] A cascade reservoir flood control storage capacity gradual release optimization dispatching system for implementing the above method comprises:
[0101] An initialization module is used to construct an optimization objective function and constraint conditions for the gradual release of the flood control storage capacity of cascade reservoirs, and generate a water level trajectory for the gradual release of the flood control storage capacity of cascade reservoirs that meets the constraint conditions based on artificial experience decision-making or conventional dynamic programming methods;
[0102] The stepwise optimization decomposition module is used to decompose the optimization scheduling problem of the stepwise release of the flood control storage capacity of the cascade reservoirs at the current iteration number into multiple two-stage stepwise optimization sub-problems;
[0103] The gradient-like deep search module performs gradient-like deep search based on the step-by-step optimization decomposition module, including gradient-like parameter generation, gradient-like state generation, state comparison and replacement, stop condition judgment and other functions to obtain a better optimization state;
[0104] The output module is used to repeatedly execute the step-by-step optimization decomposition module and the gradient-like deep search module until the preset iteration stop condition is met to obtain the optimal scheduling result, which serves as the water level optimization process for gradually releasing the flood control storage capacity of the final cascade reservoirs.
[0105] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory is used to store at least one program, and the processor is used to load the at least one program to execute the above-mentioned cascade reservoir flood control storage capacity gradual release optimization scheduling method.
[0106] The method of the present invention is further described below by means of specific examples.
[0107] Example
[0108] Take the cascade reservoir group consisting of four large reservoirs A, B, C and D in the upper reaches of a river as an example.
[0109] The engineering development task of this cascade reservoir is mainly to generate electricity, while taking into account flood control, navigation and promoting local economic and social development. The total regulating storage capacity, reserved flood control storage capacity and installed capacity of the cascade reservoir are all large, and have played a huge comprehensive benefit in flood control, power generation, navigation, ecology, and water replenishment during the dry season.
[0110] According to the joint dispatching scheme of cascade reservoirs, in real-time dispatching, when the cascade reservoirs do not need to conduct flood control dispatching in the region and do not need to cooperate with the upstream reservoirs to implement flood control dispatching, the flood control storage capacity reservation method of the cascade reservoirs can be optimized in a timely manner under the premise of ensuring flood control safety, and the flood control storage capacity of the cascade reservoirs can be gradually released. To this end, the proposed method of the present invention is applied to the optimization dispatching problem of gradually releasing the flood control storage capacity of the cascade reservoirs.
[0111] The optimal dispatching period of this embodiment is the flood season in July. During this period, the cascade reservoirs can gradually release flood control storage capacity in early, mid and late July, respectively, on the premise of meeting flood control safety and combining water supply conditions and flood control situation needs. Generally, the capacity can be 993 million m3 respectively. 3 , 2.493 billion m 3 and 2.993-6.493 billion m 3 , of which the end of July should be formulated in combination with the specific flood control situation. The following selects 15%, 30%, 50%, 70%, and 90% of the water inflow frequency in July as the wet year, the relatively wet year, the normal year, the relatively dry year, and the dry year, a total of 5 different typical water inflows, and uses the method of the present invention and POA to optimize the gradual release of the flood control storage capacity of the cascade reservoirs. The comparison of the calculation results is shown in Table 1.
[0112] Table 1: Comparison of the results of the method of the present invention and POA calculation
[0113]
[0114] From Table 1 and Figure 2 It can be seen that for different typical water inflow conditions such as a wet year, a moderately wet year, a normal year, a moderately dry year, and a dry year, the power generation calculated by the method of the present invention is similar to that of POA, but the calculation advantages are reflected in:
[0115] (1) From the comparison of power generation, the method of the present invention and POA are gradually approaching the global optimal solution, but the power generation of the method of the present invention is greater than that of POA, and the calculation accuracy is improved, indicating that after adopting the gradient-like method, the local depth search capability of the method of the present invention is better.
[0116] (2) The calculation time of the method of the present invention is significantly shorter than that of POA, which is only about 10% of the calculation time of POA. Moreover, as the calculation scale of the hydropower station increases, the calculation performance advantage becomes more prominent.
[0117] (3) By Figure 2 It can be seen that compared with the iterative calculation process of POA, the method of the present invention can quickly approach the global optimal solution in the early stage of iteration, and the search capability is very powerful.
[0118] From the above analysis, it can be seen that compared with POA, the method of the present invention can, on the one hand, increase power generation and have higher calculation accuracy, and on the other hand, save calculation time and have higher calculation efficiency. It is feasible and effective to apply the method to the optimization calculation of the gradual release of flood control storage capacity of cascade reservoirs.
[0119] Figure 3 to Figure 6The changes in the water levels of the four large reservoirs A, B, C, and D in this embodiment are listed under the conditions of normal water years, when the water levels of each reservoir gradually rise to release the flood control storage capacity. The scheduling results all meet the constraints of various water level, storage capacity, flow rate, etc. Each reservoir gradually releases the flood control storage capacity, and the power generation head and flow of each reservoir are optimized. The leading reservoir intercepts and stores water in advance to raise the reservoir water level and give play to the head and water volume effect. The water levels of other reservoirs are steadily raised, which comprehensively improves the total power generation of the cascade reservoirs and can effectively support the scheduling, operation, management and production practice of the cascade reservoirs.
[0120] In summary, the present invention first collects basic information and scheduling schemes for flood season scheduling of cascade reservoirs, constructs the objective function and constraints of the optimal scheduling of the gradual release of flood control storage capacity of cascade reservoirs, and gives the initial state of the water level process of the gradual release of flood control storage capacity of cascade reservoirs; then, based on the stepwise optimization method (POA) as the basic framework, the multi-stage cascade reservoir flood control storage capacity gradual release optimal scheduling problem is decomposed into multiple two-stage sub-problems; then, in the calculation process of each sub-problem, a gradient-like method is embedded, and a gradient-like deep search and update is performed on the initial state; finally, the global optimal solution is successively approached through iterative optimization, and the optimal scheduling process of the gradual release of flood control storage capacity of cascade reservoirs that takes into account the optimal benefits is output.
[0121] The present invention performs stage-by-stage dimensionality reduction on the optimization scheduling problem of gradually releasing the flood control storage capacity of cascade reservoirs, combines equidistant spatial parameters based on angle patterns and step control parameters based on iterative processes to generate diverse search information to seek the optimization direction of the scheduling trajectory, avoids the comprehensive combination of all discrete states of each reservoir at each stage, reduces the computational complexity, improves the local deep search capability, avoids the problem of the curse of dimensionality, and is suitable for the optimization scheduling of large-scale cascade reservoir groups.
[0122] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
[0123] Other parts not described belong to the prior art.
Claims
1. A method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs, characterized in that: The steps include: 1) Collect basic information on flood season operation of cascade reservoirs and joint operation plan of cascade reservoirs; 2) According to the joint dispatching scheme of cascade reservoirs, obtain the releasable flood control storage capacity of cascade reservoirs at each stage as the flood characteristics change; based on the basic data and the releasable flood control storage capacity, construct the objective function and constraint conditions of the gradual release of the flood control storage capacity of cascade reservoirs, and set the calculation parameters; the calculation parameters include the number of reservoirs N, the number of dispatching stages T, the maximum number of iterations M, and the maximum number of class gradient searches K; 3) Using artificial experience decision-making or conventional methods to obtain the initial state of the water level process of gradually releasing the flood control storage capacity of the cascade reservoirs that meets the constraints And calculate the discrete step length of each reservoir at each stage Δ=(Δ i,j ) N×T , the calculation formula is as follows: In the formula, and Δ i,j They represent the initial state of reservoir i at stage j respectively; and Z i,j Respectively represent the upper and lower limits of the water level of reservoir i at stage j; i = 1, 2, ..., N; j = 1, 2, ..., T; 4) Initialize the number of iterations m = 1; 5) performing a gradient-like stepwise optimization method; 6) Let m=m+1, if m>M, go to step 7); otherwise go to step 5) and continue to perform gradient-like depth search; 7) Stop the calculation and output the final optimal trajectory state, that is, obtain the optimal dispatching process of gradually releasing the flood control storage capacity of the cascade reservoirs with the best benefits.
2. The method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs according to claim 1 is characterized by: In the step 1), the basic data include characteristic parameters of each reservoir, water level storage capacity curve, downstream water level flow relationship curve, power station unit output characteristic curve, and reservoir water inflow process.
3. The method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs according to claim 1 is characterized by: In the step 2), the objective function is: Where: F is the total power generation during the gradual release of the flood control storage capacity of the cascade reservoirs; N is the number of reservoirs; i is the reservoir number, and i = 1, 2, ..., N; T is the number of stages in the scheduling period during which the flood control storage capacity of the cascade reservoirs is gradually released; j is the stage number, and j = 1, 2, ..., T; A i is the output coefficient of reservoir i; Q i,j is the power generation flow of reservoir i in stage j, m 3 / s;H i,j is the average power generation head of reservoir i in stage j after deducting the head loss, m; Δt is the stage duration, h.
4. The method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs according to claim 1 is characterized by: In step 2), the constraints include: (1) Water balance constraints of each reservoir: (2) Hydraulic connection constraints of each reservoir: (3) Water level constraints of each reservoir: (4) Constraints on outflow from each reservoir: (5) Constraints on discharge capacity of each reservoir: (6) Output constraints of each reservoir: (7) Cascade reservoirs can release flood control storage capacity constraints: (8) Constraints on total output of cascade reservoirs: (9) Non-negative constraints: All variables are non-negative; Where V i,j is the storage capacity of reservoir i at the end of stage j, m 3 ;I i,j is the inflow of reservoir i in stage j, m 3 / s;O i,j is the outflow of reservoir i in stage j, m 3 / s; and are the minimum outflow and maximum outflow of reservoir i in stage j, respectively, m 3 / s;S i,j is the water discharge of reservoir i at stage j, m 3 / s; R i,j is the interval flow of reservoir i in stage j, m 3 / s; Z i,j is the water level before the dam of reservoir i at stage j, m; and are the lowest and highest water levels in front of the dam of reservoir i at stage j, m; O i,max (Z i,j ) is the water level Z corresponding to reservoir i at stage j i,j Maximum discharge capacity, m 3 / s;P min i,j and P max i,j are the minimum and maximum output of reservoir i at stage j, kW; V i 0 is the storage capacity corresponding to the flood control limit water level of reservoir i, m 3 SV j is the total releasable flood control storage capacity of the cascade reservoirs at stage j, m 3 NP j is the minimum total output of the cascade reservoirs at stage j, kW.
5. The method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs according to claim 1 is characterized by: The step 5) comprises the following steps: 5.1) Decompose the optimization scheduling problem of gradually releasing flood control reservoir capacity into T-1 two-stage problems; 5.2) For each two-stage problem, perform gradient-like deep search in turn.
6. The method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs according to claim 5 is characterized by: The step 5.2) comprises the following steps: 5.2.1) Initialize the number of class gradient searches k = 1; 5.2.2) Status Perform gradient-like deep search to generate gradient-like states in, is the state of the sth two-stage at the mth iteration; 5.2.3) Comparison of class gradient states and current status if Better than The current state is replaced by the gradient-like state, otherwise no processing is done; 5.2.4) Let k = k + 1. If k ≤ K, it indicates that the gradient-like depth search has been completed in the current stage, and go to step 5.2.5). Otherwise, go to step 5.2.2) and continue to perform the gradient-like depth search. 5.2.5) Determine whether all two-stage sub-problems have executed gradient-like deep search. If so, go to step 6), otherwise, go to step 5.2.1) and execute gradient-like deep search on the remaining two-stage sub-problems in turn.
7. The method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs according to claim 6 is characterized by: The class gradient state The calculation formula is as follows: In the formula, α1, α2, ..., α N-1 are N-1 random angles with values in [0,2π]. According to the characteristics of trigonometric functions, the gradient-like equidistant spatial parameters D1, D2, ..., D N The value of is between -1 and 1, ensuring At the same time, g is the step size control parameter, which is set to gradually decrease with the advancement of iteration to gradually shrink the search range of the class gradient.
8. The method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs according to claim 7 is characterized by: The calculation formula of the step length control parameter g is: g=1-m / M.
9. A cascade reservoir flood control storage capacity gradual release optimization scheduling system, characterized by: include: An initialization module is used to construct an optimization objective function and constraint conditions for the gradual release of the flood control storage capacity of cascade reservoirs, and generate a water level trajectory for the gradual release of the flood control storage capacity of cascade reservoirs that meets the constraint conditions based on artificial experience decision-making or conventional dynamic programming methods; The stepwise optimization decomposition module is used to decompose the optimization scheduling problem of the stepwise release of the flood control storage capacity of the cascade reservoirs at the current iteration number into multiple two-stage stepwise optimization sub-problems; The gradient-like deep search module performs gradient-like deep search based on the step-by-step optimization decomposition module, including gradient-like parameter generation, gradient-like state generation, state comparison and replacement, and stop condition judgment functions to obtain a better optimization state; The output module is used to repeatedly execute the step-by-step optimization decomposition module and the gradient-like deep search module until the preset iteration stop condition is met to obtain the optimal scheduling result, which serves as the water level optimization process for gradually releasing the flood control storage capacity of the final cascade reservoirs.
10. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory is used to store at least one program, and the processor is used to load the at least one program to execute the method for optimizing the scheduling of the gradual release of flood control storage capacity of cascade reservoirs as described in any one of claims 1 to 8.
Citation Information
Patent Citations
A discrete gradient stepwise optimization algorithm for medium and long-term reservoir dispatching with HVDC constraints
CN109409569A
Quick solving method for multi-target short-term real-time joint scheduling of hybrid reservoir group
CN117541016A
Cascade reservoir group hydro-fluctuation scheduling variable-scale chaos search optimization method and system
CN118297167A
Method for finding solutions of optimized scheduling model for cascade hydroelectric power plant based on hybrid algorithm
JP2023178944A
Cited By
Cascade reservoir group short-term optimization scheduling method and system based on equal reservoir capacity line
CN119273118A
Dynamic reservoir capacity distribution method for improving output of large reservoir group
CN119918967A
Method for dynamic allocation of reservoir capacity for improving output of large reservoir group
CN119918967B
Cascade reservoir flood control capacity optimal distribution method considering multi-region flood control tasks
CN119962915A
Single reservoir scheduling method and device based on adaptive grid and microstep simulation
CN122047105A
