CN-122026322-A - Power system flexibility evaluation method and system for improving clustering and multivariate variation modal decomposition based on entropy weight method

CN122026322ACN 122026322 ACN122026322 ACN 122026322ACN-122026322-A

Abstract

The invention relates to the field of power system performance evaluation, and discloses a power system flexibility evaluation method and system based on entropy weight method improved clustering and multivariate variation modal decomposition, wherein the method comprises the steps of obtaining a daily payload curve and performing similarity clustering on the payload curve; decomposing daily payload curves of different clusters into payload component curves under a plurality of fluctuation frequency bands, dividing an upward climbing set and a downward climbing set to obtain upward and downward flexibility requirements under each time scale, combining the daily payload curves to obtain running states of each type of controllable unit of the power system, calculating upward and downward flexibility resources under different fluctuation time scales according to the running states, the output regulation characteristics and the fluctuation periods of each type of controllable unit, carrying out matching analysis on the flexibility resources and the flexibility requirements under the same time scale, calculating to obtain flexibility indexes of the power system, and weighting to form flexibility evaluation comprehensive indexes of the power system.

Inventors

WANG XIANBO
ZHANG ZHIYI
CHEN ZIJIE
ZHANG SIQI
YAN YUNFENG
QI DONGLIAN
CHEN YULIN

Assignees

浙江大学海南研究院
海南电网有限责任公司三亚供电局

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (10)

1. An electric power system flexibility evaluation method for improving clustering and multivariate variation modal decomposition based on an entropy weight method is characterized by comprising the following steps: Acquiring a daily payload curve, performing similarity clustering on the payload curve through a clustering algorithm improved based on an entropy weight method, and simplifying sample data into a representative typical mode; Decomposing daily payload curves of different clusters based on a multivariate variation modal decomposition algorithm, dividing the daily payload curves into payload component curves under a plurality of fluctuation frequency bands, and dividing an upward climbing set and a downward climbing set through waveform identification to obtain upward and downward flexibility requirements under each time scale; performing unit combination on daily net load curves of each typical mode to obtain the running states of each type of controllable unit of the corresponding power system, and calculating upward and downward flexible resources under different fluctuation time scales according to the running states, the output adjustment characteristics and the fluctuation periods of each type of controllable unit; And carrying out matching analysis on the flexibility resources and the flexibility demands under the same time scale, calculating the flexibility missing probability, the flexibility abundance expectation and the flexibility deficiency expectation of each time scale to serve as the flexibility index of the power system, and weighting to form the flexibility evaluation comprehensive index of the power system.
2. The power system flexibility evaluation method based on entropy weight improvement clustering and multivariate decomposition mode decomposition according to claim 1, wherein similarity clustering is performed on the payload curves through a clustering algorithm based on entropy weight improvement, and the method comprises the following steps: forming a daily net load curve by using load, wind power and photovoltaic historical output curves, and selecting daily average load, daily maximum load and daily maximum fluctuation difference as load characteristic indexes; respectively weighting different characteristic indexes by adopting an entropy weight method, and generating a new clustering center by setting an initial weight vector to improve a clustering algorithm; After generating a new clustering center, calculating the contribution degree of each characteristic index to the clustering center to obtain a similarity matrix with m evaluation objects and k evaluation indexes; updating the attraction degree matrix and the attribution degree matrix through the similarity matrix, judging whether the attraction degree matrix and the attribution degree matrix are converged or not according to the damping coefficient updating, outputting a clustering result if the attraction degree matrix and the attribution degree matrix are converged, and otherwise, updating the attraction degree matrix and the attribution degree matrix again.
3. The power system flexibility evaluation method based on entropy weight method improved clustering and multivariate decomposition mode as claimed in claim 1, wherein decomposing daily payload curves of different clusters into payload component curves in a plurality of fluctuation frequency bands based on multivariate decomposition mode decomposition algorithm comprises: determining input data according to the set IMF set decomposed by the load data ; Will input data Inputting the constructed multi-variational modal decomposition model to input data Constructing an optimal constrained variation problem; Unconstrained constraint variable problem is carried out by utilizing Lagrangian function so as to solve the variable problem, whether the solving result is converged is determined, if not, the unconstrained problem is solved by adopting an alternate direction multiplier iteration algorithm, the IMF set and the center frequency set of the IMF components are iteratively updated, a set number of IMF components are obtained so as to obtain a new IMF set, and the redetermined input data is input into a multi-variable-modal decomposition model; If the power supply control system is converged, the upward and downward flexibility requirements are determined, and the daily payload curve is divided into payload component curves of 3 frequency bands by combining the distributed power generation fluctuation characteristics and the power supply regulation rate distribution.
4. The power system flexibility evaluation method based on entropy weight method improved clustering and multivariate decomposition mode decomposition according to claim 3, wherein the up-and-down flexibility requirements under each time scale are obtained by dividing up-ramp sets and down-ramp sets through waveform identification, comprising: Splitting the net load component curves of 3 frequency bands into upward and downward 2 climbing subsets through waveform identification; Wherein, the climbing section amplitude in the climbing subset represents the flexibility requirement of the climbing section, the duration represents the fluctuation period of the climbing section, and the fluctuation period length under different fluctuation components corresponds to the time scale defined by the fluctuation period length.
5. The power system flexibility evaluation method based on entropy weight method improved clustering and multivariate decomposition mode decomposition according to claim 1, wherein the unit combination is performed on daily payload curves of each typical mode to obtain corresponding running states of each type of controllable unit of the power system, and the method comprises the following steps: The short-time flexible resource size of the thermal power station is constrained by the upward and downward climbing rates and the maximum and minimum running output of the thermal power station, and on a long-time scale, the flexibility adjustment is carried out by starting and stopping the gas unit and adjusting the output of the flexibly modified coal-fired unit besides the online running combined cycle gas unit; The hydroelectric generating set is a peak regulation power supply, and provides corresponding flexible resources according to the running state of the hydroelectric generating set in each time scale in the day; The power electronic energy storage device is a flexible power supply capable of being adjusted rapidly, wherein the super capacitor and the compressed air energy storage device are respectively suitable for flexible adjustment of a time scale of <15min and a time scale of 15-60 min; the size of the controllable load regulating resource is determined by the power load size of the demand side response implementation mechanism and the maximum output change limit of the demand side response implementation mechanism.
6. The power system flexibility evaluation method based on entropy weight method improved clustering and multivariate variation modal decomposition according to claim 1, wherein the matching analysis is performed on the flexibility resources and the flexibility demands under the same time scale, the flexibility missing probability, the flexibility abundance expectancy and the flexibility deficiency expectancy of each time scale are calculated to serve as the flexibility indexes of the power system, and the system flexibility evaluation comprehensive indexes are formed by weighting, and the method comprises the following steps: Selecting flexibility missing probability, flexibility abundance expectancy and flexibility deficiency expectancy as flexibility indexes of the power system; based on the flexibility indexes, the flexibility indexes under each scale are weighted and summed to form a scale weighted flexibility index, and the weight of each scale weighted flexibility index is determined by the average flexibility requirement under unit time so as to reflect the influence of fluctuation on the overall flexibility of the power system.
7. The method for evaluating the flexibility of the power system based on the improved clustering and the multivariate decomposition mode decomposition by the entropy weight method according to claim 6, wherein the flexibility deficiency probability, the flexibility abundance expectancy and the flexibility deficiency expectancy are selected as flexibility indexes of the power system, specifically: counting the flexibility missing climbing sections with flexibility resources lower than flexibility requirements to obtain flexibility missing probability; Setting a flexibility abundance expectation according to flexibility abundance, resources and requirements of a certain climbing section and the installed total amount of the electric power system; And calculating the sample expectation through the insufficient flexibility of the insufficient flexibility climbing section so as to reflect the severity of the power system flexibility vacancy.
8. An electric power system flexibility evaluation system based on entropy weight method improved clustering and multivariate variation modal decomposition, which is characterized by comprising: The clustering module is used for acquiring a daily payload curve, performing similarity clustering on the payload curve through a clustering algorithm improved based on an entropy weight method, and simplifying sample data into a representative typical mode; The flexibility requirement determining module is used for decomposing daily payload curves of different clusters based on a multivariate variation modal decomposition algorithm, dividing the daily payload curves into payload component curves under a plurality of fluctuation frequency bands, and dividing an upward climbing set and a downward climbing set through waveform identification to obtain upward and downward flexibility requirements under each time scale; The flexibility resource determining module is used for carrying out unit combination on the daily net load curves of each typical mode to obtain the running states of each type of controllable unit of the corresponding power system, and calculating upward and downward flexibility resources under different fluctuation time scales according to the running states, the output adjusting characteristics and the fluctuation period of each type of controllable unit; the assessment module is used for carrying out matching analysis on the flexibility resources and the flexibility demands under the same time scale, calculating the flexibility missing probability, the flexibility abundance expectation and the flexibility deficiency expectation of each time scale to be used as the flexibility index of the power system, and weighting to form the flexibility assessment comprehensive index of the power system.
9. A computer readable storage medium storing one or more programs, wherein the one or more programs comprise instructions, which when executed by a computing device, cause the computing device to perform any of the methods of claims 1-7.
10. A computing device comprising one or more processors, memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs comprising instructions for performing any of the methods of claims 1-7.

Description

Power system flexibility evaluation method and system for improving clustering and multivariate variation modal decomposition based on entropy weight method Technical Field The invention relates to the technical field of power system performance evaluation, in particular to a power system flexibility evaluation method and system based on improved clustering and multivariate variation modal decomposition by an entropy weight method. Background In recent years, with the large-scale installation grid connection of wind power, photovoltaic and other strong fluctuation and intermittent distributed energy sources, the operation of an electric power system faces new challenges. The flexibility of the power system is the capability of ensuring the reliable operation of the power system by rapidly responding to the predictable and unpredictable supply side and demand side output fluctuation of the system in the high-permeability intermittent power supply access system. Unlike conventional stabilized power supplies, the output characteristics of these distributed energy sources exhibit significant weather dependence, and their random fluctuation characteristics result in an increased fluctuation in the net load curve of the system (the actual load of the total power demand minus the output of the new energy source), which places more stringent demands on the regulation capability of conventional power supplies. Under the background, how to scientifically evaluate the capacity of the power system for adjusting the uncertainty of the supply and demand sides after the high-proportion new energy is accessed becomes an important research topic in the current energy field, and promotes the continuous innovation of a power system flexibility evaluation system and a method theory. Existing methods for quantitative assessment of power system flexibility can be divided into (1) deterministic methods and (2) uncertainty methods. The deterministic method generally adopts evaluation means such as a deterministic quantization index or a scoring table, or takes the inherent properties of elements and systems or the physical quantity obtained by simulation operation as a flexibility index. The uncertainty index mainly comprises a section type index and a probability type index, wherein the section type index measures the flexibility of the system according to the variation range of the flexibility related physical quantity. The main idea of probability shape is to describe probability distribution of physical quantities (such as renewable energy output, net load and flexibility margin) of an electric power system by using statistical parameters or mathematical transformation thereof so as to reflect the characteristics of the system flexibility in a certain aspect. The prior evaluation method has the following problems that (1) the influence of uncertainty is ignored by the determined index, the flexibility characteristic of the system cannot be comprehensively reflected, and the interval type index has a plurality of inconveniences in the aspect of describing the high distributed energy source occupation ratio. The probability index adopts low order statistics such as mean value, variance, quantile and the like to describe the flexibility of the system, can reflect the central position and discrete degree of data, but neglects high-dimensional characteristics such as distribution skewness, kurtosis and the like, and cannot clearly reflect the potential and risk of the system flexibility adjustment. (2) The method does not consider source load multi-time scale fluctuation, or the consideration of the multi-time scale is only based on the first-order difference of the payload curve, and cannot reflect the difference of flexibility requirements of different time scales. Accordingly, it is also difficult to effectively distinguish between the differences in the tuning characteristics of the system controllable unit at different time scales in terms of flexible resources. The flexibility evaluation index obtained by the method lacks the degree of distinction in the adjustment rate, which is unfavorable for the planning of a flexible power supply. Disclosure of Invention Aiming at the problems, the invention aims to provide a power system flexibility evaluation method and system for improving clustering and multivariate variation modal decomposition based on an entropy weight method, which can improve prediction stability and prediction precision. According to the technical scheme, the method comprises the steps of obtaining daily payload curves, carrying out similarity clustering on the payload curves through a clustering algorithm based on entropy weight improvement, simplifying sample data into representative typical modes, decomposing daily payload curves of different clusters based on a multi-variant modal decomposition algorithm, dividing the daily payload curves into payload component curves under a plurality of fluctuation freque