CN-121052615-B - Primary target-oriented cascade reservoir group multi-target scheduling method

CN121052615BCN 121052615 BCN121052615 BCN 121052615BCN-121052615-B

Abstract

The invention provides a primary target-oriented cascade reservoir group multi-target scheduling method, and relates to the technical field of operation and control of power systems. The method comprises the steps of determining a maximum normalized upper limit and lower limit calculation method of a scheduling target type, dividing primary targets and secondary targets, constructing an optimized scheduling model for secondary target interval building constraint, combining secondary target constraint, solving a multi-target optimal solution set, performing pareto screening, normalizing, generating an encrypted multi-target optimal non-dominant front edge through two-dimensional scattered interpolation, establishing an optimal replacement function of the secondary targets relative to the primary targets, determining importance degrees of the secondary targets, determining a secondary target constraint threshold value by combining the replacement function and the importance degrees of the secondary targets relative to the primary targets, and solving the optimized scheduling model again according to the secondary target constraint threshold value to obtain a scheduling scheme. The scheduling method solves the problems of different dimension magnitude, unknown normalization basis and insufficient emphasis of multi-attribute decision-making in multi-objective optimization.

Inventors

He zhongzheng
JI CHEN
WEI BOWEN
LI SHULIANG
LU JIAHAO
GUO JUN
GUO HAIMENG
CHEN JIAWEI

Assignees

南昌大学

Dates

Publication Date: 20260505
Application Date: 20251104

Claims (9)

1. The primary target-oriented cascade reservoir group multi-target scheduling method is characterized by comprising the following steps of: Based on the theoretical optimal values, determining the calculation modes of the upper limit and the lower limit of extremum normalization; Dividing different scheduling targets into a primary target and a secondary target, dividing target intervals of the secondary target based on the upper limit and the lower limit of extremum normalization, selecting interval constraint for the secondary target from the target intervals, constructing a secondary target constraint set, optimizing the primary target based on the secondary target constraint set, and constructing an optimized scheduling model; The method comprises the steps of obtaining an initial solution set by solving an optimized scheduling model under different space constraints, screening the initial solution set based on the pareto non-dominant theory to obtain a non-dominant solution set, normalizing the non-dominant solution set based on the extreme value normalization upper limit and the extreme value normalization lower limit to obtain a normalized non-dominant solution set, encrypting the normalized non-dominant solution set based on two-dimensional scattered interpolation to obtain a normalized non-dominant front; calculating an optimal displacement function of the secondary target relative to the primary target based on the normalized non-dominant front, determining a importance level of the secondary target relative to the primary target based on a weighting method, determining a constraint threshold of the secondary target based on the optimal displacement function and the importance level; and replacing the interval constraint in the secondary target constraint set with the constraint threshold based on single-side constraint, and solving an optimal scheduling model to obtain an optimal scheduling scheme.
2. The scheduling method according to claim 1, wherein the calculation modes of the upper limit and the lower limit of the extremum normalization are determined based on the theoretical optimal value, and the method comprises the steps that the different scheduling targets comprise a maximum scheduling target, a minimum scheduling target, an intermediate scheduling target and an interval scheduling target; fully arranging different scheduling targets, constructing multi-target scheduling models with different priority combinations, calculating the optimal solution of each multi-target scheduling model, and extracting theoretical optimal values corresponding to different scheduling targets, wherein the maximum scheduling targets are Upper limit of extremum of (2) Taking the theoretical optimal value and determining the extreme lower limit based on the guarantee rate requirement of the very large scheduling target The minimal scheduling object Lower limit of extremum of (2) Taking the theoretical optimal value and determining the extreme upper limit based on the guarantee rate requirement of the minimum scheduling target Determining the inter-type scheduling target Theoretical optimum value of (2) And determining a lower limit interval based on the guaranteed rate requirements of the intermediate scheduling targets Determining the interval type scheduling target Is the optimum interval of (2) Determining a basic interval based on the guarantee rate requirement of the interval type scheduling target ; Wherein, the At the minimum value of the lower limit interval, At the maximum value of the lower limit interval, Is the lower limit of the optimal interval, Is the upper limit of the optimal interval, Is the lower limit of the basic interval, Is the upper limit of the basic interval.
3. The scheduling method of claim 1, wherein optimizing the primary objective based on the secondary objective constraint set, when constructing the optimized scheduling model, comprises: Dividing different scheduling targets into primary targets And Secondary targets In the first step Secondary targets Upper limit of extremum of (2) And lower limit of extremum As a boundary, the secondary target Evenly discrete out Individual points Combining arbitrary adjacent two points to generate Secondary targets Target interval of (2) Generating corresponding target intervals by other secondary targets in the same way, selecting one interval from the corresponding target intervals as interval constraint for each secondary target, if the second target is Secondary targets Select the first Each target zone is taken as a zone constraint Co-formation of individual secondary targets The secondary target constraint set is constructed as follows: ; the security constraints include equality constraints Inequality constraint And carrying out single-objective optimization on the primary objective based on the safety constraint and the secondary objective constraint set to construct an optimized scheduling model, wherein the optimized scheduling model is expressed as follows: 。
4. the scheduling method of claim 3, wherein obtaining the normalized non-dominant solution set comprises based on a target interval number of the secondary targets Determination of Discrete accuracies of the individual secondary targets, based on which are generated by permutation and combination The method comprises the steps of obtaining target constraint combinations, solving an optimized scheduling model under each target constraint combination, screening feasible solutions to obtain an initial solution set, screening the initial solution set based on the pareto non-dominant theory to obtain a non-dominant solution set, and carrying out extremum normalization processing on the non-dominant solution set based on the upper limit and the lower limit of extremum normalization to obtain a normalized non-dominant solution set.
5. The scheduling method of claim 1, wherein encrypting the normalized non-dominant solution set based on two-dimensional scattered interpolation to obtain a normalized non-dominant front comprises, based on two-dimensional scattered interpolation, within a normalized preset range, according to interpolation accuracy Encrypting the normalized non-dominant solution set to obtain a result Normalized non-dominant front composed of interpolation points : ; ; Wherein, the Representing normalized non-dominant front The first of (3) The number of normalized non-dominant solutions is, 、 Respectively represent Primary target value of medium normalization Normalized secondary target value, subscript Represent the first Secondary targets, subscripts Indicating that the target value belongs to the first The normalized non-dominant solution.
6. The scheduling method of claim 5, wherein calculating an optimal permutation function of the secondary target relative to the primary target based on the normalized non-dominant front comprises defining the optimal permutation function as an optimal delta of the primary target that the unit secondary target can trade for; The calculation process comprises the steps of selecting a normalized non-dominant solution of the normalized non-dominant front as a calculation point, traversing all secondary targets of the calculation point to obtain adjacent non-dominant solutions meeting preset difference conditions, calculating increment estimated values of the secondary targets relative to a primary target based on a difference method and the adjacent non-dominant solutions, taking the maximum value in the calculation result as an optimal increment, and determining an optimal replacement function, wherein the formula is as follows: ; Wherein, the Representing secondary targets Relative primary objective Is used for the optimal permutation function of (c), Indicating the accuracy of the interpolation, And Representing normalized secondary target values Is used to determine the adjacent non-dominant solution of (c), Representing a normalized primary target value, wherein the preset difference condition comprises The differential method comprises a left differential, a right differential and a center differential.
7. The scheduling method of claim 6, wherein determining a constraint threshold for a secondary objective based on the optimal permutation function and the importance level comprises the importance level including absolute advantage, equalization, disadvantage, and absolute disadvantage; Based on secondary objectives Relative to the primary target Determines the importance level of the secondary targets respectively Relative to the primary target Is the optimal permutation function of (2) Threshold of (2) And when the importance level is absolute advantage , And when the importance level is dominant , And when the importance level degree is balanced , And when the importance level is inferior , And when the importance level is absolute disadvantage , ; Based on the relation between the threshold and the optimal displacement function, the back-extrapolation obtains the condition Secondary objectives of (2) Based on the determination method of the extreme upper limit and the extreme lower limit, obtaining a secondary target And a final constraint threshold.
8. The scheduling method of claim 7, wherein replacing the interval constraint in the secondary target constraint set with the constraint threshold based on the single-side constraint, and solving an optimal scheduling model to obtain an optimal scheduling scheme, comprises determining a single-side boundary based on the constraint threshold of the single-side constraint on the secondary target, replacing the interval constraint in the secondary target constraint set with the processed constraint threshold, and solving an updated constraint optimal scheduling model to obtain the optimal scheduling scheme, wherein the updated constraint optimal scheduling model is as follows: 。
9. The scheduling method of claim 1, wherein the weighting method comprises a subjective weighting method and an objective weighting method, and wherein the two-dimensional random interpolation comprises a radial basis function interpolation and an inverse distance weighted interpolation.

Description

Primary target-oriented cascade reservoir group multi-target scheduling method Technical Field The invention relates to the technical field of operation and control of power systems, in particular to a primary target-oriented cascade reservoir group multi-target scheduling method. Background In the comprehensive utilization of water resources, a cascade reservoir group is used as a core infrastructure for guaranteeing water supply, power generation and ecological demands, and multi-objective collaborative scheduling still faces a plurality of key technical bottlenecks, and is mainly characterized in that firstly, the difference of dimension and magnitude of each scheduling target is obvious, index units such as generated energy, water supply, ecological flow and the like are different, the numerical span is large, the optimization result is unbalanced, secondly, the target normalization basis is ambiguous, the definition of the scheduling target type is ambiguous, the subjectivity of the normalization result is strong, the reliability is insufficient, thirdly, the multi-attribute decision is insufficient, so that the scheduling scheme tends to be conserved due to the fact that part of targets are too tightly constrained, or the potential risk is included due to the fact that part of targets are excessively compromised. The traditional scheduling method is difficult to systematically overcome the problems, and the comprehensive benefit of the cascade reservoir group is restricted to be fully exerted. It is therefore desirable to provide a solution to the above-mentioned problems. Disclosure of Invention The invention aims to provide a primary object-oriented cascade reservoir group multi-object scheduling method which can solve the problems of unbalanced object dimension magnitude, deficient normalization basis and insufficient decision mechanism in the existing scheduling method. The invention provides a primary object-oriented cascade reservoir group multi-object scheduling method, which comprises the following steps: Based on the theoretical optimal values, determining the calculation modes of the upper limit and the lower limit of extremum normalization; Dividing different scheduling targets into a primary target and a secondary target, dividing target intervals of the secondary target based on the upper limit and the lower limit of extremum normalization, selecting interval constraint for the secondary target from the target intervals, constructing a secondary target constraint set, optimizing the primary target based on the secondary target constraint set, and constructing an optimized scheduling model; The method comprises the steps of obtaining an initial solution set by solving an optimized scheduling model under different space constraints, screening the initial solution set based on the pareto non-dominant theory to obtain a non-dominant solution set, normalizing the non-dominant solution set based on the extreme value normalization upper limit and the extreme value normalization lower limit to obtain a normalized non-dominant solution set, encrypting the normalized non-dominant solution set based on two-dimensional scattered interpolation to obtain a normalized non-dominant front; calculating an optimal displacement function of the secondary target relative to the primary target based on the normalized non-dominant front, determining a importance level of the secondary target relative to the primary target based on a weighting method, determining a constraint threshold of the secondary target based on the optimal displacement function and the importance level; and replacing the interval constraint in the secondary target constraint set with the constraint threshold based on single-side constraint, and solving an optimal scheduling model to obtain an optimal scheduling scheme. According to the cascade reservoir group multi-target scheduling method for the primary targets, an optimal scheduling model is built by defining different scheduling targets, the upper limit and the lower limit of the extreme value of the optimal scheduling targets and the primary and secondary priorities of the scheduling targets, and a constraint set is redetermined based on a defined optimal replacement function, so that the optimal scheduling model is solved, and an optimal scheduling scheme is obtained. Optionally, when determining the calculation modes of the upper limit and the lower limit of the extremum normalization based on the theoretical optimal value, the method comprises the steps that different scheduling targets comprise a very large scheduling target, a very small scheduling target, an intermediate scheduling target and an interval type scheduling target; fully arranging different scheduling targets, constructing multi-target scheduling models with different priority combinations, calculating the optimal solution of each multi-target scheduling model, and extracting theoretical optimal values corresponding