Successive approximation method of float-discrete differential dynamic programming for cascade reservoir operation

Through the float-discrete differential dynamic programming successive approximation method, the water level operation trajectory of the cascade reservoir group is optimized, and the problems of calculation time-consuming and local optimal solutions in the existing technology are solved, and more efficient cascade reservoir group scheduling is achieved.

CN118395058BActive Publication Date: 2024-11-26WUHAN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410567854.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-08
Publication Date
2024-11-26
Estimated Expiration
2044-05-08

Smart Images

  • Figure CN118395058B_ABST
    Figure CN118395058B_ABST
Patent Text Reader

Abstract

The present invention provides a float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling, which belongs to the technical field of reservoir scheduling, and includes the following steps: raising the water level of the cascade reservoir group in order from small to large according to the water surface area of ​​the reservoir; the cascade reservoir group after the water level is raised generates electricity according to the incoming water; at the end of the calculation period, the water level of the cascade reservoir group after the water level is raised falls back to the set water level, and the initial water level operation trajectory is obtained; based on the obtained initial water level operation trajectory, with the maximum cascade total power generation target, discrete differential dynamic programming calculations are performed from upstream to downstream reservoirs to obtain the improved water level operation trajectory of each cascade reservoir; the improved water level operation trajectory of each cascade reservoir is used as the initial trajectory, and iterative optimization is performed. The present invention makes full use of the relationship between the output and water head of the cascade hydropower station, reduces the invalid calculation workload, and improves the calculation efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of reservoir regulation, in particular to a float-discrete differential dynamic programming successive approximation method for cascade reservoir group regulation. Background Art

[0002] The essence of the optimal dispatching of cascade hydropower stations is the establishment and solution of the model. In terms of model establishment, for the medium- and long-term dispatching of cascade hydropower stations, the optimal dispatching model with the maximum power generation as the goal is the basic guarantee for the full utilization of the cascade hydropower resources in the basin. In terms of the solution of this model, the most commonly used solution methods are the dynamic programming successive approximation (DPSA) algorithm and the intelligent algorithm.

[0003] The DPSA algorithm is based on dynamic programming. It performs dynamic programming calculations on each reservoir from upstream to downstream to obtain the initial trajectory. Then, the operation trajectory of other reservoirs is fixed and the dynamic programming method is used to calculate them one by one from top to bottom. The benefit value at this time is the total benefit of the cascade. This cycle is repeated until the operation trajectory of each reservoir remains unchanged or the total benefit of the cascade converges. The intelligent algorithm is mainly based on genetic algorithm and particle swarm algorithm, and uses massive calculations to obtain better target values.

[0004] The DPSA algorithm can obtain a better optimal solution, but each reservoir is calculated several times using the dynamic programming method, which results in a large amount of computational workload and a long time consumption. Intelligent algorithms are prone to fall into local optimal solutions, and the solutions obtained are unstable, making it difficult to obtain global optimal solutions. These solution methods all rely too much on mathematical methods, without taking into account the changing laws of water energy in cascade reservoirs, resulting in a large number of calculations being useless. Based on this, the present invention proposes a float-discrete differential dynamic programming successive approximation method for the scheduling of cascade reservoir groups, which proposes new ideas and methods for solving the optimal scheduling model of cascade reservoir groups. Summary of the invention

[0005] The purpose of the present invention is to provide a float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling, which fully utilizes the relationship between the output and water head of the cascade hydropower station, reduces invalid calculation workload, and improves calculation efficiency.

[0006] To achieve the above object, the present invention provides a float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling, comprising the following steps:

[0007] S1. Raise the water level of the cascade reservoirs in order of their water surface area from small to large;

[0008] S2, after the water level is raised, the cascade reservoirs generate electricity according to the incoming water;

[0009] S3. At the end of the calculation period, the water level of the cascade reservoir group after the water level is raised falls back to the set water level, and the initial water level operation trajectory is obtained;

[0010] S4, based on the initial water level operation trajectory obtained, with the maximum total cascade power generation target, discrete differential dynamic programming calculations are performed from upstream to downstream reservoirs to obtain improved water level operation trajectories of each cascade reservoir;

[0011] S5. Taking the improved water level operation trajectory of each cascade reservoir as the initial trajectory, iterative optimization is performed according to step S4 until the water level operation trajectory of each cascade reservoir remains unchanged or the total power generation value of the cascade calculation period remains unchanged.

[0012] Preferably, in step S1, the water level of the cascade reservoirs is increased in order from small to large according to the water surface area of ​​the reservoirs. The specific operation is:

[0013] S101, obtaining the water surface area of ​​each reservoir according to the water level-area relationship curve of each reservoir based on the given initial water level of each reservoir;

[0014] S102. At the beginning of the calculation period, the last reservoir discharges water to generate electricity according to the ecological flow requirements, and other reservoirs generate electricity according to the expected output power generation mode. The water storage capacity of each reservoir is adjusted so that the reservoir with the smallest water surface area is stored first until it is full.

[0015] S103, storing water in the order of water surface area from small to large until all reservoirs are full.

[0016] Preferably, in step S3, the water level of the cascade reservoir group after the water level is raised falls back to the set water level, and an initial water level operation trajectory is obtained, and the specific operations are:

[0017] S301, starting from the given water level at the end of the last period of the first reservoir, according to the predicted natural water inflow, and in reverse order according to the expected power output, calculate the initial water level of the period;

[0018] S302, based on the initial water level of the time period calculated in step S301, the initial water level of the previous time period is calculated in reverse order according to the expected power generation;

[0019] S303, repeating step S302 until the initial water level of a certain period exceeds the water level calculated in step S1, at which time, the output value of this period is recalculated according to the water level calculated in step S1, and the water level change process of the first hydropower station is obtained by combining it with the water level calculated in step S1;

[0020] S304. For the cascade reservoirs, repeat steps S301 to S303 in order from upstream to downstream until the calculation of the most downstream reservoir is completed, and the initial water level operation trajectory of each cascade reservoir is obtained.

[0021] Preferably, in step S4, based on the obtained initial water level operation trajectory, discrete differential dynamic programming calculations are performed for each reservoir from upstream to downstream to obtain an improved water level operation trajectory for each cascade reservoir, and the specific operations are as follows:

[0022] From the initial water level operation trajectory of the first reservoir, 2 to 3 discrete points are taken at a certain step length above and below the water level value at the end of each period (except the last period) to form a corridor, and the initial water level operation trajectory of other reservoirs is fixed. With the goal of maximizing the total power generation of the cascade during the calculation period, dynamic programming calculation is performed in the corridor to obtain a new water level operation trajectory; based on the new water level operation trajectory, the corridor is reselected for dynamic programming calculation until the new water level operation trajectory is stable, and the discrete step length is reduced and then the dynamic programming calculation is performed again until the discrete step length meets the preset accuracy and the water level operation trajectory is stable;

[0023] According to the above steps, the calculations for the second and third reservoirs are performed respectively until the calculation of the last reservoir is completed, and the improved water level operation trajectory of each cascade reservoir is obtained;

[0024] Based on the improved water level operation trajectory of the cascade reservoirs, the above process is repeated until the improved water level operation trajectory of the cascade reservoirs no longer changes.

[0025] Preferably, the maximum objective function of the total cascade power generation is:

[0026]

[0027] Where E is the cascade power generation, kWh; I is the number of hydropower stations; i is the sequence number of the hydropower station, i∈[1,I]; T is the number of dispatch periods; t is the sequence number of the dispatch period, t∈[1,T]; K it is the output coefficient of the ith hydropower station in the tth period; N it is the average output of the i-th hydropower station in the t-th period, kW; Q it is the average power generation flow of the i-th hydropower station in the t-th period, m 3 / s;H it is the average generating head of the ith hydropower station in the tth period, m; ΔT t is the length of the tth period, h.

[0028] Preferably, the optimal dispatching model of cascade hydropower stations with the goal of maximizing total power generation satisfies the following constraints:

[0029] Reservoir water level constraints:

[0030]

[0031] Among them, Z i (t+1) is the water level of the i-th reservoir at the end of period t, m; is the lower limit of the reservoir water level at the end of period t of the i-th reservoir, m; is the upper limit of the reservoir water level at the end of period t of the i-th reservoir, m;

[0032] Water balance constraints:

[0033]

[0034] Among them, V i (t+1) and V i (t) are the storage capacity of the i-th reservoir at the end of period t and the beginning of period t, respectively, m 3 ; is the inflow of the i-th reservoir at the beginning of period t, m 3 / s; is the outflow of the i-th reservoir at the beginning of period t, m 3 / s; is the loss flow of the i-th reservoir at the beginning of period t, m 3 / s;

[0035] Hydropower station output constraints:

[0036]

[0037] Among them, P i (t) is the output of the i-th reservoir in period t, kW; is the lower limit of the output of the i-th reservoir in period t, kW; is the upper limit of the output of the i-th reservoir in period t, kW;

[0038] Downflow flow restriction:

[0039]

[0040] in, is the lower limit of the outflow of the i-th reservoir in period t, which is controlled by the ecological flow demand or the power generation flow corresponding to the minimum output, m 3 / s; is the maximum discharge flow of the i-th reservoir in period t, which is controlled by the discharge capacity, m 3 / s;

[0041]

[0042] in, is the inflow of the i+1th reservoir during period t, m 3 / s; is the interval flow between the i-th reservoir and the i+1-th reservoir in period t, m 3 / s;

[0043] Final water level constraint:

[0044]

[0045] Among them, Z i (T+1) is the final water level of the i-th reservoir in period T, m; Calculate the water level at the end of the period for the i-th reservoir, m;

[0046] Non-negativity constraint: All variables must be non-negative.

[0047] Therefore, the present invention adopts the above-mentioned float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling, fully utilizes the relationship between the output and water head of the cascade hydropower station, reduces invalid calculation workload, and improves calculation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling of the present invention;

[0049] Figure 2 Flow chart for raising the water level of cascade reservoirs;

[0050] Figure 3 A flow chart to obtain the initial water level trajectory of each cascade reservoir. DETAILED DESCRIPTION

[0051] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0052] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0053] Embodiment 1

[0054] like Figure 1 As shown, it is a flow chart of a float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling of the present invention, which includes the following steps:

[0055] S1. Raise the water level of the cascade reservoirs in order from small to large according to the water surface area of ​​the reservoirs. The specific operation is as follows (e.g. Figure 2 shown):

[0056] S101, obtaining the water surface area of ​​each reservoir according to the water level-area relationship curve of each reservoir based on the given initial water level of each reservoir;

[0057] S102. At the beginning of the calculation period, the last reservoir discharges water for power generation according to the ecological flow requirements, and other reservoirs generate electricity according to the expected power generation mode (generating electricity as large as possible without discarding water). The water storage capacity of each reservoir is adjusted so that the reservoir with the smallest water surface area is stored first until it is full.

[0058] S103, store water in order from small to large water surface area until all reservoirs are full

[0059] S2, after the water level is raised, the cascade reservoirs generate electricity according to the incoming water;

[0060] When the cascade reservoirs are full, the water head of the cascade hydropower stations is at its maximum. At this time, power generation is carried out according to the incoming water, and the power generation efficiency of each cascade hydropower station is relatively high.

[0061] S3. At the end of the calculation period, the water level of the cascade reservoir group after the water level is raised falls back to the set water level, and the initial water level operation trajectory is obtained. The specific operation is as follows (such as Figure 3 shown):

[0062] S301, starting from the given water level at the end of the last period of the first reservoir, according to the predicted natural water inflow, and in reverse order according to the expected power output, calculate the initial water level of the period;

[0063] S302, based on the initial water level of the time period calculated in step S301, the initial water level of the previous time period is calculated in reverse order according to the expected power generation;

[0064] S303, repeating step S302 until the initial water level of a certain period exceeds the water level calculated in step S1, at which time, the output value of this period is recalculated according to the water level calculated in step S1, and the water level change process and outflow flow process of the first hydropower station are obtained by combining it with the water level calculated in step S1;

[0065] S304. For the cascade reservoirs, repeat steps S301 to S303 in order from upstream to downstream until the calculation of the most downstream reservoir is completed, and the initial water level operation trajectory of each cascade reservoir is obtained.

[0066] S4. Based on the initial water level operation trajectory obtained, discrete differential dynamic programming calculations are performed from upstream to downstream reservoirs to obtain improved water level operation trajectories of each cascade reservoir. The specific operations are as follows:

[0067] From the initial water level operation trajectory of the first reservoir, 2 to 3 discrete points are taken at a certain step length above and below the water level value at the end of each period (except the last period) to form a corridor, and the initial water level operation trajectory of other reservoirs is fixed. With the goal of maximizing the total power generation of the cascade during the calculation period, dynamic programming calculation is performed in the corridor to obtain a new water level operation trajectory; based on the new water level operation trajectory, the corridor is reselected for dynamic programming calculation until the new water level operation trajectory is stable, and the discrete step length is reduced and then the dynamic programming calculation is performed again until the discrete step length meets the preset accuracy and the water level operation trajectory is stable;

[0068] According to the above steps, the calculations for the second and third reservoirs are performed respectively until the calculation of the last reservoir is completed, and the improved water level operation trajectory of each cascade reservoir is obtained;

[0069] Based on the improved water level operation trajectory of the cascade reservoirs, the above process is repeated until the improved water level operation trajectory of the cascade reservoirs no longer changes.

[0070] The maximum objective function of the total cascade power generation is:

[0071]

[0072] Where E is the cascade power generation, kWh; I is the number of hydropower stations; i is the sequence number of the hydropower station, i∈[1,I]; T is the number of dispatch periods; t is the sequence number of the dispatch period, t∈[1,T]; K it is the output coefficient of the ith hydropower station in the tth period; N it is the average output of the i-th hydropower station in the t-th period, kW; Q it is the average power generation flow of the i-th hydropower station in the t-th period, m 3 / s;H it is the average generating head of the ith hydropower station in the tth period, m; ΔT t is the length of the tth period, h.

[0073] The optimal dispatching model of cascade hydropower stations with the goal of maximizing total power generation meets the following constraints:

[0074] Reservoir water level constraints:

[0075]

[0076] Among them, Z i (t+1) is the water level of the i-th reservoir at the end of period t; is the lower limit of the reservoir water level at the end of period t of the i-th reservoir; is the upper limit of the reservoir water level at the end of period t of the i-th reservoir;

[0077] Water balance constraints:

[0078]

[0079] Among them, V i (t+1) and V i (t) are the storage capacity of the i-th reservoir at the end of period t and at the beginning of period t; is the inflow of the i-th reservoir at the beginning of period t; is the outflow of the i-th reservoir at the beginning of period t; is the loss flow of the i-th reservoir at the beginning of period t;

[0080] Hydropower station output constraints:

[0081]

[0082] Among them, P i (t) is the output of the i-th reservoir in period t; is the lower limit of the output of the i-th reservoir in period t; is the upper limit of the output of the i-th reservoir in period t;

[0083] Downflow flow restriction:

[0084]

[0085] in, is the lower limit of the outflow of the i-th reservoir in period t, which is controlled by the ecological flow demand or the power generation flow corresponding to the minimum output; is the maximum discharge flow of the i-th reservoir in period t, which is controlled by the discharge capacity;

[0086]

[0087] in, is the inflow of the i+1th reservoir during period t; is the interval flow between the i-th reservoir and the i+1-th reservoir in period t;

[0088] Final water level constraint:

[0089]

[0090] Among them, Z i (T+1) is the final water level of the i-th reservoir in period T; Calculate the water level at the end of the period for the i-th reservoir;

[0091] Non-negativity constraint: All variables must be non-negative.

[0092] Therefore, the present invention adopts the above-mentioned float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling, fully utilizes the relationship between the output and water head of the cascade hydropower station, reduces invalid calculation workload, and improves calculation efficiency.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling, characterized in that: The following steps are involved: S1. Raise the water level of the cascade reservoirs in order of their water surface area from small to large; S2, after the water level is raised, the cascade reservoirs generate electricity according to the incoming water; S3. At the end of the calculation period, the water level of the cascade reservoir group after the water level is raised falls back to the set water level, and the initial water level operation trajectory is obtained; S4, based on the initial water level operation trajectory obtained, with the maximum total cascade power generation target, discrete differential dynamic programming calculations are performed from upstream to downstream reservoirs to obtain improved water level operation trajectories of each cascade reservoir; S5, taking the improved water level operation trajectory of each cascade reservoir as the initial trajectory, and iterating and optimizing according to step S4 until the water level operation trajectory of each cascade reservoir remains unchanged or the total power generation value of the cascade calculation period remains unchanged; In step S4, based on the initial water level operation trajectory obtained, with the maximum total cascade power generation target, discrete differential dynamic programming calculations are performed from upstream to downstream reservoirs to obtain improved water level operation trajectories of each cascade reservoir. The specific operations are as follows: From the initial water level trajectory of the first reservoir, 2 to 3 discrete points are taken above and below the water level value at the end of each period to form a corridor. The initial water level trajectory of other reservoirs is fixed. With the goal of maximizing the total power generation of the cascade during the calculation period, dynamic programming calculation is performed in the corridor to obtain a new water level trajectory. Based on the new water level trajectory, the corridor is reselected for dynamic programming calculation until the new water level trajectory is stable. Then, the discrete steps are reduced and dynamic programming calculation is performed again until the discrete steps meet the preset accuracy and the water level trajectory is stable. According to the above steps, the calculations for the second and third reservoirs are performed respectively until the calculation of the last reservoir is completed, and the improved water level operation trajectory of each cascade reservoir is obtained; Based on the improved water level operation trajectory of the cascade reservoirs, the above process is repeated until the improved water level operation trajectory of the cascade reservoirs no longer changes; The optimal dispatching model of cascade hydropower stations with the goal of maximizing total power generation meets the following constraints: Reservoir water level constraints: Among them, Z i (t+1) is the water level of the i-th reservoir at the end of period t, m; is the lower limit of the reservoir water level at the end of period t of the i-th reservoir, m; is the upper limit of the reservoir water level at the end of period t of the i-th reservoir, m; Water balance constraints: Among them, V i (t+1) and V i (t) are the storage capacity of the i-th reservoir at the end of period t and the beginning of period t, respectively, m 3 ; is the inflow of the i-th reservoir at the beginning of period t, m 3 / s; is the outflow of the i-th reservoir at the beginning of period t, m 3 / s; is the loss flow of the i-th reservoir at the beginning of period t, m 3 / s; Hydropower station output constraints: Among them, P i (t) is the output of the i-th reservoir in period t, kW; is the lower limit of the output of the i-th reservoir in period t, kW; is the upper limit of the output of the i-th reservoir in period t, kW; Downflow flow restriction: in, is the lower limit of the outflow of the i-th reservoir in period t, which is controlled by the ecological flow demand or the power generation flow corresponding to the minimum output, m 3 / s; is the maximum discharge flow of the i-th reservoir in period t, which is controlled by the discharge capacity, m 3 / s; Water flow continuity constraints: in, is the inflow of the i+1th reservoir during period t, m 3 / s; is the interval flow between the i-th reservoir and the i+1-th reservoir in period t, m 3 / s; Final water level constraint: Among them, Z i (T+1) is the final water level of the i-th reservoir in period T, m; Calculate the water level at the end of the period for the i-th reservoir, m; Non-negativity constraint: All variables must be non-negative.

2. The method of successive approximation of float-discrete differential dynamic programming for cascade reservoir group regulation according to claim 1 is characterized in that: In step S1, the water level of the cascade reservoirs is increased in order from small to large according to the water surface area of ​​the reservoirs. The specific operation is: S101, obtaining the water surface area of ​​each reservoir according to the water level-area relationship curve of each reservoir based on the given initial water level of each reservoir; S102. At the beginning of the calculation period, the last reservoir discharges water to generate electricity according to the ecological flow requirements, and other reservoirs generate electricity according to the expected output power generation mode. The water storage capacity of each reservoir is adjusted so that the reservoir with the smallest water surface area is stored first until it is full. S103, storing water in the order of water surface area from small to large until all reservoirs are full.

3. The float-discrete differential dynamic programming successive approximation method for cascade reservoir group scheduling according to claim 2 is characterized in that: In step S3, the water level of the cascade reservoir group after the water level is raised falls back to the set water level, and the initial water level operation trajectory is obtained. The specific operations are: S301, starting from the given water level at the end of the last period of the first reservoir, according to the predicted natural water inflow, and in reverse order according to the expected power output, calculate the initial water level of the period; S302, based on the initial water level of the time period calculated in step S301, the initial water level of the previous time period is calculated in reverse order according to the expected power generation; S303, repeating step S302 until the initial water level of a certain period exceeds the water level of step S1, at which time, the output value of this period is recalculated according to the water level of step S1, and the water level change process of the first hydropower station is obtained by combining it with the water level of step S1; S304. For the cascade reservoirs, repeat steps S301 to S303 in order from upstream to downstream until the calculation of the most downstream reservoir is completed, and the initial water level operation trajectory of each cascade reservoir is obtained.

4. The float-discrete differential dynamic programming successive approximation method for cascade reservoir group regulation according to claim 3 is characterized in that: The maximum objective function of the total cascade power generation is: Where E is the cascade power generation; I is the number of hydropower stations; i is the serial number of the hydropower station, i∈[1,I]; T is the number of dispatching periods; t is the serial number of the dispatching period, t∈[1,T]; K it is the output coefficient of the ith hydropower station in the tth period; N it is the average output of the i-th hydropower station in the t-th period; Q it is the average power generation flow of the i-th hydropower station in the t-th period; H it is the average generating head of the ith hydropower station in the tth period; ΔT t is the duration of the tth period.