CN112766564B - Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group - Google Patents

Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group Download PDF

Info

Publication number
CN112766564B
CN112766564B CN202110054612.9A CN202110054612A CN112766564B CN 112766564 B CN112766564 B CN 112766564B CN 202110054612 A CN202110054612 A CN 202110054612A CN 112766564 B CN112766564 B CN 112766564B
Authority
CN
China
Prior art keywords
hydropower station
stage
discrete
hydropower
scheduling
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110054612.9A
Other languages
Chinese (zh)
Other versions
CN112766564A (en
Inventor
邹强
饶光辉
何小聪
喻杉
柳林云
胡学东
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Changjiang Institute of Survey Planning Design and Research Co Ltd
Original Assignee
Changjiang Institute of Survey Planning Design and Research Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Changjiang Institute of Survey Planning Design and Research Co Ltd filed Critical Changjiang Institute of Survey Planning Design and Research Co Ltd
Priority to CN202110054612.9A priority Critical patent/CN112766564B/en
Publication of CN112766564A publication Critical patent/CN112766564A/en
Application granted granted Critical
Publication of CN112766564B publication Critical patent/CN112766564B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • General Health & Medical Sciences (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention relates to the technical field of efficient utilization of water resources and optimal scheduling of hydropower, and discloses a step hydropower station group optimal scheduling stage-by-stage reverse learning dimension reduction optimization method, which comprises the steps of firstly calculating to obtain an initial scheduling state process and discrete step length of each hydropower station; then for each stage of calculation, randomly generating discrete state numbers of the upper side and the lower side on the upper side and the lower side of the initial state combination stage by stage and combining the discrete state numbers to form a corridor, and calculating the reverse state combination of the initial state to carry out iterative optimization on the basis of the upper boundary and the lower boundary of the corridor so as to obtain an improved scheduling process; repeating the above process until all stages are calculated; and finally, shrinking the discrete step length, and repeating iteration until convergence so as to approach to a global optimal solution and output an optimal scheduling process. The step hydropower station group optimization scheduling phase-by-phase reverse learning dimensionality reduction optimization method effectively reduces the calculation complexity, greatly improves the calculation efficiency, and is suitable for large-scale step hydropower station group optimization scheduling.

Description

Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group
Technical Field
The invention relates to the technical field of efficient water resource utilization and hydropower optimization scheduling, in particular to a step hydropower station group optimization scheduling stage-by-stage reverse learning dimension reduction optimization method.
Background
In recent years, with the rapid and rapid development and rapid and orderly promotion of water conservancy and hydropower industry in China, the scale of hydropower stations is increasing day by day, and particularly, in hydropower bases in extra large watersheds such as Jinshajiang, Yamo and Wujiang, an unprecedented super-large-scale hydropower system in the world is formed. The optimal scheduling of the cascade hydropower station group has remarkable social, economic and ecological comprehensive benefits, can effectively promote the efficient utilization of the hydropower resources in the drainage basin, improve the scheduling management level of the hydropower station group and ensure the safe, stable and reliable operation of a power grid system. However, the increasing number of hydropower stations and the gradual increase of system scale directly lead to exponential increase of calculation time and occupied memory in the optimization scheduling process, the contradiction between the complexity of the problem and the limitation of the solving method is obvious, the dimension disaster problem becomes a scientific problem which is bound to be faced by the optimization scheduling of hydropower stations, and the direct relation is to clean power supply and effective utilization of energy in China.
The cascade hydropower station group has complex hydrology, water power and electric power relation and comprises a plurality of constraint conditions such as water level, reservoir capacity, flow and output, the reservoir station group optimal scheduling is a complex coupling optimal control problem with large scale, high dimensionality, multiple stages, strong constraint and nonlinearity, and effective and efficient solution faces a larger technical bottleneck.
When the traditional method is applied to the problem of the cascade hydropower station group combined optimization scheduling, along with the increase of the number of power stations and the discrete number, a severe dimension disaster problem is faced.
Mathematically, the optimization scheduling problem of the hydropower station group is a large-scale, high-dimensional, multi-stage, strong-constraint and nonlinear optimization problem. To solve the problem, researchers at home and abroad successively put forward various methods such as linear programming, nonlinear programming, Dynamic Programming (DP), Discrete Differential Dynamic Programming (DDDP), stepwise optimization (POA), evolutionary algorithm and the like. Although the methods are widely applied to the field of hydropower station group optimal scheduling, the defects of premature convergence, high calculation cost, dimension disaster and the like still exist when the large-scale hydropower station group optimal scheduling problem is processed. Therefore, development of novel and effective optimization methods is urgently needed, so that synchronous reduction of memory occupation and calculation time consumption is realized, the problem of dimension disaster is effectively relieved, and the calculation efficiency and the solving precision of large-scale hydropower system combined scheduling are ensured.
Reverse learning is a new technology proposed in the field of computational intelligence in recent years, and the core idea of the reverse learning is to evaluate the current state and the reverse state at the same time and use preferentially, so that the search process is accelerated.
Disclosure of Invention
The invention aims to provide a step-by-step reverse learning dimensionality reduction optimization method for optimization scheduling of a cascade hydropower station group, aiming at the defects of the technology, and the method converts a high-dimensional hydropower optimization scheduling problem into a relatively simple low-dimensional optimization sub-problem, avoids comprehensive combination among discrete states of each stage of each power station, and has quick solving efficiency and calculation precision.
In order to achieve the purpose, the step hydropower station group optimization scheduling phase-by-phase reverse learning dimension reduction optimization method comprises the following steps:
1) determining initial calculation conditions including an objective function, constraint conditions and decision variables of the optimal scheduling of the cascade hydropower station group;
2) setting calculation parameters including maximum iteration times M, total stage number T, maximum reverse learning times F and convergence precision epsilon;
3) setting initial test tracks of all hydropower stations according to a conventional dynamic planning method or a manual experience decision
Figure BDA0002900453200000021
And calculating and acquiring initial discrete step length delta (delta) of each hydropower stationi,j)N×T，
Figure BDA0002900453200000022
Wherein the content of the first and second substances,
Figure BDA0002900453200000023
representing the initial state, Δ, of the hydroelectric station i in phase ji,jRepresenting a hydroelectric power planti at the initial discrete step size of stage j,
Figure BDA0002900453200000024
representing the upper water level limit of the hydropower station i in stage j, i,jZrepresenting the lower limit of the water level of the hydropower station i in the stage j, wherein N is the number of the hydropower stations, T is the number of stages of the dispatching period of the hydropower stations, and K is the initial discrete number;
4) setting the iteration number m to be 1;
5) setting the phase j to 1;
6) forming a corridor and constructing reverse state combinations of each stage by the current state of each hydropower station, randomly generated discrete numbers of the upper side and the lower side and discrete step length, wherein in the stage j, the f-th reverse learning, the discrete number of the upper side of the hydropower station i is
Figure BDA0002900453200000031
Discrete number of lower side of
Figure BDA0002900453200000032
Corresponding upper gallery boundary
Figure BDA0002900453200000033
Lower gallery boundary
Figure BDA0002900453200000034
According to the calculation principle of reverse learning, in the optimized corridor range S1,S2]Initial state of hydropower station i in inner stage j
Figure BDA0002900453200000035
In the reverse state of
Figure BDA0002900453200000036
Comparison Zi,jAnd
Figure BDA0002900453200000037
if Z isi,jIs superior to
Figure BDA0002900453200000038
Then let Zi,jReplacement of
Figure BDA0002900453200000039
Repeating the reverse learning discrete differential dynamic programming method until the maximum reverse learning frequency F of the stage is reached;
7) j is j +1, if j is more than T, the step 8) is carried out, and if j is more than T, the step 6) is carried out;
8) if it is
Figure BDA00029004532000000310
Shrinking discrete step length of all hydropower station states to order
Figure BDA00029004532000000311
Go to step 9), otherwise go to step 5);
9) let M be M +1, if M > M or
Figure BDA00029004532000000312
Go to step 10), otherwise go to step 5).
10) Stopping calculation and outputting the final optimal track Z0。
Compared with the prior art, the invention has the following advantages:
1. the method is based on the reverse learning idea, the reverse state combination of the initial state is calculated in the generated corridor stage by stage to carry out iterative optimization until convergence, and the method is visual, simple and convenient and easy to operate;
2. the high-dimensional hydropower optimization scheduling problem is converted into a relatively simple low-dimensional optimization sub-problem, comprehensive combination among all discrete states of each stage of each power station is avoided, the optimization direction of the track is searched through reverse learning, the calculation complexity is effectively reduced, and the calculation efficiency is greatly improved;
3. the calculation result of the method is reasonable and reliable, and the method can serve management operation and production practice of the cascade hydropower station group.
Drawings
FIG. 1 is a flow chart of a step hydropower station group optimization scheduling phase-by-phase reverse learning dimension reduction optimization method of the invention;
FIG. 2 is a graph comparing the power generation variation of the method of the present invention and the DDDP method at different horizontal years;
FIG. 3 is a schematic diagram of reservoir level and output variation processes of the Uedo hydropower station under the condition of water in the open water year;
FIG. 4 is a schematic diagram of the reservoir water level and output change process of a white crane beach hydropower station under the condition of the water of the open water year;
FIG. 5 is a schematic diagram of reservoir water level and output change of a brook ferry hydropower station under water conditions in an open water year;
fig. 6 is a schematic diagram of reservoir water level and output change process to a dam hydropower station under open water conditions.
Detailed Description
The invention is described in further detail below with reference to the figures and the specific embodiments.
As shown in fig. 1, a stepwise hydropower station group optimization scheduling phase-by-phase reverse learning dimension reduction optimization method includes the following steps:
1) determining initial computing conditions including an objective function, constraint conditions and decision variables of the cascade hydropower station group optimization scheduling, wherein the cascade hydropower station group optimization scheduling model can be described as follows: the initial water level and the final water level of each reservoir in the dispatching period, the warehousing runoff process and the interval runoff process are known, under the condition that various complex constraints such as the corresponding water level, flow and output of a hydropower station group are met, the optimal stage water level operation process of the hydropower station group is determined, the water abandon is reduced as far as possible while the average water head of power generation is increased, the water energy resource is utilized to the maximum extent, the maximum total generated energy of the hydropower station group in the dispatching period is realized, and the objective function of the cascade hydropower station group optimization dispatching mathematical model is as follows:
Figure BDA0002900453200000041
in the formula: f is the total generated energy in the dispatching period, N is the number of the hydropower stations, i is the serial number of the hydropower stations, and i is 1,2, …, N, T is the total number of stages in the dispatching period, j is the serial number of the stages, and j is 1,2, …, T; a. theiThe output coefficient of the reservoir i is;Qi,jfor the generated flow (m) of reservoir i in stage j3/s)，Hi,jThe method is characterized in that an average generating head (m) of head loss is deducted from a stage j for a reservoir i, delta j is a stage length (h), and a large number of complex constraint conditions need to be considered for a hydropower scheduling problem to ensure feasibility and usability of an optimization result, and the method mainly comprises the following steps:
hydraulic connection constraint:
Figure BDA0002900453200000051
water balance constraint:
Figure BDA0002900453200000052
time interval water level constraint:
Figure BDA0002900453200000053
fourthly, restricting the discharge:
Figure BDA0002900453200000054
power station output restraint:
Figure BDA0002900453200000055
sixthly, restraining the initial and final water levels: z is a linear or branched memberi,start＝Zi,end
And (c) system output constraint:
Figure BDA0002900453200000056
-non-negative constraints: each variable is non-negative;
the decision variable is water level;
in the formula: i isi,jWarehousing traffic (m) for hydropower station i at stage j3/s)，Ri,jFor the section inflow (m) of the hydropower station i in phase j3/s)，Si-1,jReject flow (m) at stage j for the i-1 st hydropower station3/s)，Vi,jFor the value of the reservoir capacity (m) of the hydropower station i at the end of the phase j3)，Zmin i,jAnd Zmax i,jThe lowest dam front water level and the highest dam front water level value (m), Q of the hydropower station i in the stage j are respectivelymin i,jAnd Qmax i,jMinimum generated flow (m) of hydropower station i in stage j3S) and maximum generated flow (m)3/s)，Pmin i,jAnd Pmax i,jMinimum power output (kW) and maximum power output (kW), Z of hydropower station i in stage j respectivelyi,startAnd Zi,endThe initial reservoir level (m) and the final control level (m, NP) of the dispatching period of the hydropower station ijIs the minimum output (kW) of the system in stage j;
2) setting calculation parameters including maximum iteration times M, total stage number T, maximum reverse learning times F and convergence precision epsilon;
3) setting initial test tracks of all hydropower stations according to a conventional dynamic planning method or a manual experience decision
Figure BDA0002900453200000061
And calculating and obtaining the initial discrete step length delta (delta) of each hydropower stationi,j)N×T，
Figure BDA0002900453200000062
Wherein, the first and the second end of the pipe are connected with each other,
Figure BDA0002900453200000063
representing the initial state, Δ, of the hydroelectric station i in phase ji,jRepresenting the initial discrete step size of the hydropower station i in phase j,
Figure BDA0002900453200000064
representing the upper water level limit of the hydropower station i in stage j, i,jZrepresenting the lower limit of the water level of the hydropower station i in the stage j, wherein N is the number of the hydropower stations, T is the number of stages of the dispatching period of the hydropower stations, and K is the initial discrete number;
4) setting the iteration number m to be 1;
5) setting the phase j to 1;
6) randomly generating discrete number and discrete steps of upper side and lower side according to current state of each hydropower stationLong, forming galleries and constructing combinations of reversal states for each stage, at stage j, the f-th reversal learning, with discrete number of the upper side of the hydroelectric station i as
Figure BDA0002900453200000065
A discrete number of the lower side of
Figure BDA0002900453200000066
Corresponding upper gallery boundary
Figure BDA0002900453200000067
Lower gallery boundary
Figure BDA0002900453200000068
According to the calculation principle of reverse learning, in the range of the optimized corridor [ S ]1,S2]Initial state of hydropower station i in inner stage j
Figure BDA0002900453200000069
In the reverse state of
Figure BDA00029004532000000610
Comparison Zi,jAnd
Figure BDA00029004532000000611
if Z isi,jIs superior to
Figure BDA00029004532000000612
Then let Zi,jReplacement of
Figure BDA00029004532000000613
Repeating the reverse learning discrete differential dynamic programming method until the maximum reverse learning frequency F of the stage is reached;
7) j is j +1, if j is more than T, the step 8) is carried out, and if j is more than T, the step 6) is carried out;
8) if it is
Figure BDA00029004532000000614
Shrinking discrete step length of all hydropower station states to order
Figure BDA00029004532000000615
Go to step 9), otherwise go to step 5);
9) let M be M +1, if M > M or
Figure BDA00029004532000000616
Go to step 10), otherwise go to step 5).
10) Stopping calculation and outputting the final optimal track Z0。
The optimized scheduling problem of the downstream cascade hydropower station group of the Jinshajiang river in China is taken as an example for research. The drainage basin has abundant water energy resources, and is a main base for strategic water resource areas and hydropower development of water resource allocation in China. The Jinshajiang river downstream cascade planning comprises Wudong De, white Crane beach, stream Luo Du and four large reservoirs facing to a family dam, the regulation performance is strong, the hydraulic connection is tight, and the total installed capacity of 4 hydropower stations is up to 46400 MW. 3 different incoming waters of dry year (frequency 75%), open water year (frequency 50%) and rich water year (frequency 25%) are selected, and the method disclosed by the invention and DDDP are respectively adopted to carry out the combined optimized scheduling of the cascade hydropower station group.
Table 1 lists the comparison of the method of the invention and the DDDP calculation taking into account different numbers of plants, different horizontal years.
TABLE 1 comparison of the method of the invention and DDDP calculation results
Figure BDA0002900453200000071
FIG. 2 is a comparison graph of the variation process of the power generation amount of two methods at different levels, and as can be seen from FIG. 2 and Table 1, the method of the present invention can obtain the power generation amount similar to DDDP, but the calculation advantages are reflected in that: firstly, the calculation time of the method is obviously shorter than that of DDDP and is only about 2 percent of that of DDDP, and the calculation performance advantage is more prominent along with the increase of the calculation scale of the hydropower station; secondly, the method and the DDDP are close to the global optimal solution, but the method can obtain the generating capacity similar to that of the DDDP, meanwhile, the calculation efficiency is greatly improved, and the larger the calculation scale is, the more remarkable the advantages are.
Fig. 3, 4, 5 and 6 show schematic diagrams of the change processes of the reservoir water level and the output force of udon, white beach, stream ferry, hydropower station to the home dam respectively. The optimization results of all hydropower stations meet various constraint conditions, the water level gradually disappears in the dry period, and the water level is quickly lifted in the water storage period, so that the scale benefit and the cascade compensation benefit of the hydropower station group are fully exerted, and the characteristic of 'enrichment and withering' is reflected, which shows that the calculation result of the method is reasonable and reliable, and can be used in management operation and production practice of the cascade hydropower station group.
Therefore, compared with the existing DDDP, the method can greatly reduce the calculation time, obviously improve the operation efficiency, and can be used for optimized dispatching of the giant step hydropower station group.
The invention fully analyzes the problem of the united optimization scheduling of cascade hydropower stations, excavates the problem of dimension disaster faced by the traditional method, and firstly calculates to obtain the initial scheduling state process and discrete step length of each hydropower station; then for each stage of calculation, randomly generating discrete state numbers of the upper side and the lower side on the upper side and the lower side of the initial state combination stage by stage and combining the discrete state numbers to form a corridor, and calculating the reverse state combination of the initial state to carry out iterative optimization on the basis of the upper boundary and the lower boundary of the corridor so as to obtain an improved scheduling process; repeating the above steps until all the stages are calculated; and finally, shrinking the discrete step length, and repeating iteration until convergence, so as to approach to a global optimal solution and output an optimal scheduling process. Aiming at the problem of the cascade hydropower station group joint optimization scheduling, the most common maximum generated energy of the cascade hydropower station group joint optimization scheduling problem is adopted as an optimization scheduling target to carry out specific research so as to more remarkably improve the benefit of the cascade hydropower station group joint optimization scheduling.
The method converts the high-dimensional hydropower optimization scheduling problem into a relatively simple low-dimensional optimization sub-problem, avoids comprehensive combination of all discrete states of each stage of each power station, effectively reduces the calculation complexity by searching the optimization direction of the track by combining reverse learning, greatly improves the calculation efficiency, solves the dimension disaster problem existing in the traditional method during the optimization scheduling solving of the complex hydropower system, has quick solving efficiency and calculation precision, and is suitable for the optimization scheduling of large-scale cascade hydropower stations.
The invention is based on the principle of reverse learning, maps the information of each stage of the initial state of the scheduling process to the reverse state for space search, eliminates the degraded dimensional information, can increase the search range of the state space information, reduces the high dimensional information interference among the stages, and improves the search efficiency and the calculation precision. In other words, the method disclosed by the invention is innovative in that the optimization track is searched through reverse learning, so that the comprehensive combination of discrete states of all stages of all hydropower stations is avoided, the calculation complexity is effectively reduced, and the method has good support and application values for the research of the optimization scheduling problem of a large-scale hydropower system.

Claims (1)

1. A step hydropower station group optimization scheduling phase-by-phase reverse learning dimension reduction optimization method is characterized by comprising the following steps of: the method comprises the following steps:
1) determining initial calculation conditions, including an objective function, constraint conditions and decision variables of the cascade hydropower station group optimization scheduling;
2) setting calculation parameters including maximum iteration number M, total number T of stages, maximum reverse learning number F and convergence precision epsilon;
3) setting initial test tracks of all hydropower stations according to a conventional dynamic planning method or artificial experience decision
Figure FDA0002900453190000011
And calculating and obtaining the initial discrete step length delta (delta) of each hydropower stationi,j)N×T，
Figure FDA0002900453190000012
Wherein the content of the first and second substances,
Figure FDA0002900453190000013
representing the initial state, Δ, of the hydroelectric station i in phase ji,jRepresenting the initial discrete step size of the hydropower station i in phase j,
Figure FDA0002900453190000014
representing the upper water level limit of the hydropower station i in stage j, i,jZrepresenting the lower limit of the water level of the hydropower station i in the stage j, wherein N is the number of the hydropower stations, T is the number of stages of the dispatching period of the hydropower stations, and K is the initial discrete number;
4) setting the iteration number m to be 1;
5) setting the phase j to 1;
6) forming a corridor and constructing reverse state combinations of each stage by the current state of each hydropower station, randomly generated discrete numbers of the upper side and the lower side and discrete step length, wherein in the stage j, the f-th reverse learning, the discrete number of the upper side of the hydropower station i is
Figure FDA0002900453190000015
Discrete number of lower side of
Figure FDA0002900453190000016
Corresponding upper gallery boundary
Figure FDA0002900453190000017
Lower gallery boundary
Figure FDA0002900453190000018
According to the calculation principle of reverse learning, in the optimized corridor range S1,S2]Initial state of hydropower station i in inner stage j
Figure FDA0002900453190000019
In the reverse state of
Figure FDA00029004531900000110
Comparison Zi,jAnd
Figure FDA00029004531900000111
if Z isi,jIs superior to
Figure FDA00029004531900000112
Then let Zi,jReplacement of
Figure FDA00029004531900000113
Repeating the reverse learning discrete differential dynamic programming method until the maximum reverse learning frequency F of the stage is reached;
7) j is j +1, if j is more than T, the step 8) is carried out, and if j is more than T, the step 6) is carried out;
8) if it is
Figure FDA0002900453190000021
Shrinking discrete step length of all hydropower station states, and order
Figure FDA0002900453190000022
Go to step 9), otherwise go to step 5);
9) let M equal to M +1 if M > M or
Figure FDA0002900453190000023
Go to step 10), otherwise go to step 5);
10) stopping calculation and outputting the final optimal track Z0。
CN202110054612.9A 2021-01-15 2021-01-15 Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group Active CN112766564B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110054612.9A CN112766564B (en) 2021-01-15 2021-01-15 Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110054612.9A CN112766564B (en) 2021-01-15 2021-01-15 Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group

Publications (2)

Publication Number Publication Date
CN112766564A CN112766564A (en) 2021-05-07
CN112766564B true CN112766564B (en) 2022-06-03

Family

ID=75701826

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110054612.9A Active CN112766564B (en) 2021-01-15 2021-01-15 Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group

Country Status (1)

Country Link
CN (1) CN112766564B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104166887A (en) * 2014-08-21 2014-11-26 大连理工大学 Orthogonal discrete differential dynamic programming method for cascade hydropower station group joint optimization scheduling
CN106981888A (en) * 2017-05-10 2017-07-25 西安理工大学 The multiple target dynamic dispatching method of Thermal and Hydroelectric Power Systems is stored based on the complementary wind of multi-source
CN108764545A (en) * 2018-05-17 2018-11-06 湖北工业大学 A kind of Hydro-Thermal Systems short-term economic dispatching method based on social affection's particle cluster algorithm
CN109948842A (en) * 2019-03-12 2019-06-28 大连理工大学 A kind of Long-term Hydropower Station Scheduling method controlling reservoir level variation frequency
CN110751365A (en) * 2019-09-11 2020-02-04 华中科技大学 Multi-target balanced scheduling method and system for cascade reservoir group

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9373960B2 (en) * 2013-03-13 2016-06-21 Oracle International Corporation Computerized system and method for distributed energy resource scheduling

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104166887A (en) * 2014-08-21 2014-11-26 大连理工大学 Orthogonal discrete differential dynamic programming method for cascade hydropower station group joint optimization scheduling
CN106981888A (en) * 2017-05-10 2017-07-25 西安理工大学 The multiple target dynamic dispatching method of Thermal and Hydroelectric Power Systems is stored based on the complementary wind of multi-source
CN108764545A (en) * 2018-05-17 2018-11-06 湖北工业大学 A kind of Hydro-Thermal Systems short-term economic dispatching method based on social affection's particle cluster algorithm
CN109948842A (en) * 2019-03-12 2019-06-28 大连理工大学 A kind of Long-term Hydropower Station Scheduling method controlling reservoir level variation frequency
CN110751365A (en) * 2019-09-11 2020-02-04 华中科技大学 Multi-target balanced scheduling method and system for cascade reservoir group

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Forecasting Monthly Runoff Time Series by Single-Layer Feedforward Artificial Neural Network and Grey Wolf Optimizer;Xiong Cheng;《IEEE》;20200826;全文 *
基于改进鲸鱼算法的三峡水电站经济运行模型及其应用;杨堃;《水电能源科学》;20200630;全文 *
基于聚集度自适应反向学习粒子群算法在水库优化调度中的应用;邓志诚;《水利水电技术》;20200430;全文 *

Also Published As

Publication number Publication date
CN112766564A (en) 2021-05-07

Similar Documents

Publication Publication Date Title
CN106951985B (en) Multi-objective optimal scheduling method for cascade reservoir based on improved artificial bee colony algorithm
CN108711892B (en) Optimal scheduling method of multi-energy complementary power generation system
CN102855591A (en) Method and system for optimizing scheduling for short-term combined generation of cascade reservoir group
CN110838733B (en) Photovoltaic capacity configuration method suitable for cascade water-light complementary energy power generation system
CN109636043B (en) Adaptive optimization method and system for power generation dispatching of cascade hydropower system
Bai et al. Multi-objective optimal operation model of cascade reservoirs and its application on water and sediment regulation
CN109919350B (en) Multi-stage substation expansion planning method considering overall economy
CN104166887B (en) Orthogonal discrete differential dynamic programming method for cascade hydropower station group joint optimization scheduling
CN114386760A (en) Step hydropower station energy storage scheduling method considering photovoltaic access condition
CN104504455B (en) A kind of lower GROUP OF HYDROPOWER STATIONS Long-term Optimal Dispatch method of step accumulation of energy control
CN105427052A (en) Reference line-based parallel reservoir certainty optimized dispatching method
CN112700080A (en) Multistage optimal scheduling method for cascade hydropower
CN112431772A (en) Optimal dispatching operation method for flood prevention and drainage pump station group in polder region
CN115438852A (en) Practical solution method for short-term peak regulation of cascade hydropower station group
Xu et al. Developments and characteristics of pumped storage power station in China
CN112766564B (en) Stepwise reverse learning dimensionality reduction optimization method for optimization scheduling of cascade hydropower station group
CN111476477A (en) Power generation benefit target-based medium and long term optimization scheduling method for cascade hydropower station
CN115907402B (en) Method and system for evaluating joint guaranteed output of cascade hydropower station
CN106292298B (en) Series-parallel cascade pumping station water-carriage system space-time hierarchy optimization operation method and system
CN116776496A (en) Pumped storage transformation method for axial flow turbine runner
CN108520317B (en) Latin square dynamic planning and dimension reduction method for hydropower system joint optimization scheduling
CN112801816B (en) Resource optimization scheduling method for total benefits of wind, light and water complementary system
CN112766562B (en) Cascade reservoir group ecological scheduling optimization method considering historical track knowledge
CN111027825B (en) Energy storage scheduling graph drawing method based on typical dry year and output coefficient optimization
CN204140145U (en) Based on variable frequency pump and the siphonal once-through cycle water system of inverted U

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant