CN111859283A - Scene generation method considering multi-energy charge-time sequence and correlation - Google Patents

Scene generation method considering multi-energy charge-time sequence and correlation Download PDF

Info

Publication number
CN111859283A
CN111859283A CN202010487888.1A CN202010487888A CN111859283A CN 111859283 A CN111859283 A CN 111859283A CN 202010487888 A CN202010487888 A CN 202010487888A CN 111859283 A CN111859283 A CN 111859283A
Authority
CN
China
Prior art keywords
correlation
scene
matrix
energy
time sequence
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.)
Granted
Application number
CN202010487888.1A
Other languages
Chinese (zh)
Other versions
CN111859283B (en
Inventor
王丹
雷杨
李家熙
王培汀
李思源
黄德裕
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.)
Tianjin University
Original Assignee
Tianjin University
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 Tianjin University filed Critical Tianjin University
Priority to CN202010487888.1A priority Critical patent/CN111859283B/en
Publication of CN111859283A publication Critical patent/CN111859283A/en
Application granted granted Critical
Publication of CN111859283B publication Critical patent/CN111859283B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/25—Design optimisation, verification or simulation using particle-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/04—Power grid distribution networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Mathematical Analysis (AREA)
  • Software Systems (AREA)
  • Mathematical Optimization (AREA)
  • Economics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Pure & Applied Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Geometry (AREA)
  • Computational Mathematics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Computer Hardware Design (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Algebra (AREA)
  • Operations Research (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses a scene generation method considering multi-energy charge-time sequence and correlation, which comprises the following steps: 1) carrying out scene data normalization processing according to historical data; 2) based on the data after normalization processing, determining the scene clustering number by adopting a Davies-Bouldin index, and carrying out scene clustering by utilizing a K-means method; 3) establishing a probability distribution model of the multi-energy load, and fitting distribution parameters by adopting a maximum likelihood estimation method according to the established probability distribution model; 4) analyzing time sequence autocorrelation characteristics under different scenes, and generating a scene with autocorrelation by adopting an LHS-CD method; 5) extracting scenes from scene clusters with autocorrelation under different source load categories, combining to generate a scene set, and calculating the cross correlation of the scene set; 6) and (3) screening out a cross-correlation scene set which is close to the cross-correlation with the historical data on an MATLAB simulation platform by adopting a particle swarm optimization to obtain a final multi-energy load scene set with time sequence and correlation.

Description

Scene generation method considering multi-energy charge-time sequence and correlation
Technical Field
The invention relates to the field of comprehensive energy and scene generation, in particular to a scene generation method considering multi-energy charge-time sequence and correlation.
Background
With the complexity of high-permeability renewable power supply access and distribution network user power utilization behaviors, the randomness and volatility of multi-energy loads provide great challenges for operation, planning and scheduling of a power system. The construction of the uncertainty model of renewable energy and load is the key to realize the safe and stable operation, economic dispatch, reasonable planning of renewable energy and the like of the comprehensive energy system. The scene analysis method can analyze source and load uncertainty according to a scene set of potential multi-energy load, provides decision basis for work such as scheduling and planning, and reduces negative influence of the source and load uncertainty. Existing scene generation methods can be divided into probabilistic model generation methods, classical scene generation methods, and deep learning generation methods. The probability model generation method is to generate wind power or load scenes according to statistical experience or probability distribution by combining a probability model method with a Monte Carlo sampling method. The classical scene generation method usually adopts a data mining technology and a scene reduction technology to reduce or optimize a large-scale historical scene, and generates a group of classical scenes representing the whole region to be solved. The deep learning generation method based on the deep learning framework can deeply mine data, deeply analyze the statistical rules in the data and realize the unsupervised generation of scenes. At present, the research on the uncertainty model of multiple energy sources and multiple loads is sufficient, and various scenario generation methods are provided for the operation planning of an energy system. However, the current scene generation method still has room for improvement. On the one hand, the random nature of the multiple energy sources, multiple loads is determined by their inherent type characteristics. On the other hand, their fluctuation levels periodically change with time and environmental changes. Therefore, the scene generation method is constructed by considering the time sequence, the autocorrelation and the cross correlation of the scene, and a foundation is provided for the follow-up research.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides a scene generation method considering multi-energy charge-time sequence and correlation, which comprises the following steps: carrying out scene data normalization processing according to historical data; determining the number of scene clusters by using a DBI index, and carrying out scene clustering by using a K-means method; establishing a probability distribution model of the multi-energy load, and fitting distribution parameters by adopting a maximum likelihood estimation method according to the probability distribution model; analyzing time sequence autocorrelation characteristics under different scenes, and generating a scene with autocorrelation by adopting an LHS-CD method; extracting scenes from scene clusters with autocorrelation under different source load categories, combining to generate a scene set, and calculating the cross correlation of the scene set; and screening out a cross-correlation scene set which is relatively close to the cross-correlation of the historical data by adopting a particle swarm algorithm to obtain a multi-energy charge scene set with time sequence and correlation finally.
The purpose of the invention is realized by the following technical scheme:
a scene generation method considering multi-energy charge-time sequence and correlation comprises the following steps:
1) carrying out scene data normalization processing according to historical data; based on the data after normalization processing, determining the scene clustering number by adopting a Davies-Bouldin index, and carrying out scene clustering by utilizing a K-means method;
2) Establishing a probability distribution model of the multi-energy load, and fitting distribution parameters by adopting a maximum likelihood estimation method according to the established probability distribution model;
3) analyzing time sequence autocorrelation characteristics under different scenes, and generating a scene with autocorrelation by adopting an LHS-CD method; extracting scenes from scene clusters with autocorrelation under different source load categories, combining to generate a scene set, and calculating the cross correlation of the scene set;
4) and (3) screening out a cross-correlation scene set which is close to the cross-correlation with the historical data on an MATLAB simulation platform by adopting a particle swarm optimization to obtain a final multi-energy load scene set with time sequence and correlation.
Further, the Davies-Bouldin index is specifically as follows:
Figure BDA0002519778190000021
in the formula
Figure BDA0002519778190000022
Representing the distance from the ith cluster sample to the cluster center;
Figure BDA0002519778190000023
representing the distance from the jth cluster sample to the cluster center; NUMjClustering number of the jth cluster samples; w is aiRepresenting the ith cluster centroid; w is ajRepresenting the jth cluster centroid; i Wi-wj||2Representing the distance between the centroids of the ith cluster and the jth cluster; the Davies-Bouldin index reflects the mean value of the maximum similarity in each class, and the smaller the value, the better the clustering effect.
Further, a formula involved in analyzing the time sequence autocorrelation characteristics under different scenes is specifically as follows:
Figure BDA0002519778190000024
Figure BDA0002519778190000025
Wherein s isiIs the ith scene value;
Figure BDA0002519778190000026
a clustering center for the kth seed charge; r isi,i+gIs the element of the ith row and the ith + g column of the autocorrelation matrix; τ is the number of time sequences;ρautois an autocorrelation matrix.
Further, the probability distribution model of the multi-energy load is specifically as follows:
Figure BDA0002519778190000031
in the formula
Figure BDA0002519778190000032
A mathematical expectation representing k energy loads at time t;
Figure BDA0002519778190000033
the variance of k energy loads at t moment is represented;
Figure BDA0002519778190000034
the load value of the kth load at the time t is shown;
Figure BDA0002519778190000035
is the upper limit of the load value;
Figure BDA0002519778190000036
is the lower limit of the load value;
Figure BDA0002519778190000037
is the probability density function of the k load at the time t.
Furthermore, the source in the multi-energy load is renewable energy, and comprises wind power and photovoltaic power; the probability distribution model of wind power and photovoltaic specifically comprises the following steps:
Figure BDA0002519778190000038
wherein v represents wind speed; k represents a shape parameter of the Weibull distribution; c represents a scale parameter of the Weibull distribution; f (v) a probability density function representing wind speed.
Figure BDA0002519778190000039
Wherein S represents the intensity of light; srRepresenting a rated illumination intensity; α, β represent distribution parameters of the beta distribution; f (S) represents a probability density function of the illumination intensity.
Further, the LHS-CD method specifically comprises the following steps:
assume m random variables X with correlation1,X2,…,Xi,…,XmAnd XiHas a probability distribution of Yi＝F(Xi)；
1) And (3) sampling matrix X extraction: n sample points are extracted for each random variable, at the moment, an Nxm-order sampling matrix is extracted, and in order to ensure that the extracted samples uniformly cover the whole adopted interval, the random variable X is used iThe probability distribution function is divided into m subintervals, the probability corresponding to each interval is equal, and a sampling value is extracted from each interval; to avoid direct processing of the probability distribution function, Y is addedi＝F(Xi) The longitudinal axis of the method is divided into N parts averagely, N is the average number of the LHS method, the length of each interval is 1/N, one sample point is randomly taken from N intervals which are not intersected with each other, and then Y is used for obtaining the average number of the LHS methodiThe inverse function of (A) to (B) to obtain a random variable XiOf the sampled value, thus the random variable XiThe nth sample value of (d) is:
Figure BDA00025197781900000310
forming N sampling values of each random variable into a row of a matrix to obtain an Nxm-order matrix;
2) randomly generating a sequence matrix L, and calculating a Spearman correlation coefficient matrix rho of the L by adopting a time sequence autocorrelation characteristic formulaLResolving and eliminating the correlation generated by random arrangement by adopting Cholesky; target correlation coefficient matrix ρ for historical dataobjCholesky decomposition is carried out to enable random arrangement of the sequence matrix L to generate corresponding correlation; updating the initial sample matrix X to obtain a new sample matrix XuMake the arrangement order of each row of elements and GuThe element sequence of corresponding row in the sequence table is the same, and the obtained XuHaving a target relevance;
ρL＝QQT
G＝Q-1L
ρobj＝PPT
Gu＝PG＝PQ-1L
where rhoLA correlation matrix which is an order matrix L; q is a decomposed diagonal matrix; qTA transposed matrix that is Q; q -1An inverse matrix of Q; g is a transition matrix; rhoobjIs a target correlation matrix; p is a decomposed diagonal matrix; guSamples with a specified target correlation.
Further, the cross-correlation analysis method specifically comprises the following steps:
for two time sequential scenarios
Figure BDA0002519778190000041
And
Figure BDA0002519778190000042
the cross-correlation matrix is:
Figure BDA0002519778190000043
Figure BDA0002519778190000044
wherein M is the number of source charge variables; si,sjTime sequence scene values of the ith and jth moments;
Figure BDA0002519778190000045
a clustering center for the kth seed charge;
Figure BDA0002519778190000046
a clustering center of the u-th seed and load; ri,jIs the cross-correlation coefficient; rhocrossIs a cross correlation matrix.
Compared with the prior art, the technical scheme of the invention has the following beneficial effects:
1. by applying the scene generation method, a scene generation method considering multi-energy charge characteristics, time sequence, autocorrelation and cross correlation can be constructed. And (3) constructing a time sequence probability distribution model of the multi-energy loads by analyzing the distribution characteristics in the historical data, and generating a scene. And analyzing the autocorrelation characteristic and the cross-correlation characteristic of the historical data, and screening the scenes to obtain the final scene. The generated scene has the characteristics of historical data, the dimensionality of the historical data can be enriched, scene simulation data is formed, and data support is provided for links of planning, operation, transaction, evaluation and the like of the comprehensive energy system. The accuracy of research and analysis is improved, and the problems of data errors, huge data quantity, data loss and the like are solved.
2. The method comprises the steps of analyzing historical data, modeling wind speed, illumination, electric load, gas load and thermal load by considering the cross correlation among source loads and the time sequence and time sequence autocorrelation of each element, and generating a scene. Establishing a model of wind speed, illumination and multi-energy load, adopting a K-means algorithm to perform a clustering method on a plurality of original scenes, adopting an LHS-CD (long-term evolution-compact) technology to generate a series of scenes with certain time sequence autocorrelation, and utilizing a particle swarm optimization method to perform scene screening. Scene data more suitable for follow-up research and analysis can be obtained.
3. The scene method can provide help for detailed medium-term prediction or energy operation scene construction, and accurately depict the power generation characteristics of local renewable energy and the behavior habit of user energy consumption. The utilization efficiency of energy can be improved, and the planning design of a comprehensive energy system and the operation and maintenance of equipment can be better served.
Drawings
FIG. 1 illustrates two steps of a scene generation method considering multi-energy charge-time sequence and correlation;
FIG. 2 shows two steps of a scene generation method considering multi-energy charge-time sequence and correlation;
FIG. 3 shows the results of clustering and time-series distribution parameter analysis;
FIG. 4 is a diagram illustrating comparison of results of different scene generation methods;
FIG. 5 is a schematic diagram showing comparison of evaluation indexes of different scene generation methods;
Detailed Description
The invention is described in further detail below with reference to the figures and specific examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The embodiment of the invention firstly constructs a probability distribution model of renewable energy sources and multi-energy loads. The renewable energy mainly considers photovoltaic and wind power, and the multi-energy load considers electric power, gas and heat load. In the embodiment, a truncated normal distribution model is used for describing the uncertainty of the multi-energy load, a beta distribution is used for describing the uncertainty of the photovoltaic, and a Weibull distribution is used for describing the uncertainty of the wind power. Referring to fig. 1 and 2, the specific steps are as follows:
1. multi-energy load model construction
The probability distribution model of the multi-energy load is specifically as follows:
Figure BDA0002519778190000061
in the formula
Figure BDA0002519778190000062
A mathematical expectation representing k energy loads at time t;
Figure BDA0002519778190000063
the variance of k energy loads at t moment is represented;
Figure BDA0002519778190000064
the load value of the kth load at the time t is shown;
Figure BDA0002519778190000065
is the upper limit of the load value;
Figure BDA0002519778190000066
is the lower limit of the load value;
Figure BDA0002519778190000067
is the probability density function of the k load at the time t.
The probability distribution model of wind power and photovoltaic specifically comprises the following steps:
Figure BDA0002519778190000068
Figure BDA0002519778190000069
in the formula vWTRepresenting wind speed; k represents a shape parameter of the Weibull distribution; c represents a scale parameter of the Weibull distribution; v. ofin,vout,vrRespectively cut-in wind speed, cut-out wind speed and rated wind speed. Pr,PWTGRespectively representing rated power and wind power output power. f (v)WT) As a function of the probability density of the wind speed.
Figure BDA00025197781900000610
Figure BDA00025197781900000611
Wherein S represents the intensity of light; srRepresenting a rated illumination intensity; α, β represent distribution parameters of the beta distribution. PPVG,r,PPVGRespectively photovoltaic rated power and photovoltaic output power. f (S) is a probability density function of the illumination intensity.
2. Correlation analysis modeling
(1) Auto-correlation properties
For a time sequential scenario
Figure BDA00025197781900000612
The autocorrelation of the time k and the time g is shown as follows:
Figure BDA0002519778190000071
Figure BDA0002519778190000072
in the formula, ri,i+gThe autocorrelation coefficient is the autocorrelation coefficient of the time i and the time i + g; rhoautoIs an autocorrelation matrix; siIs a time sequence scene value;
Figure BDA0002519778190000073
a clustering center for the kth seed charge; τ is the number of epochs.
(2) Cross correlation property model
For two time sequential scenarios
Figure BDA0002519778190000074
And
Figure BDA0002519778190000075
the calculation formula of the cross correlation matrix is as follows:
Figure BDA0002519778190000076
Figure BDA0002519778190000077
in the formula, Ri,jThe cross correlation coefficient of the variable i and the variable j is obtained; rhocrossIs a cross-correlation matrix; siAnd sjTime sequence scene values of i moment and j moment;
Figure BDA0002519778190000078
and
Figure BDA0002519778190000079
a cluster center for the kth or u seed charge; m is the number of source charge variables.
3. Scene generation method considering multi-energy charge-time sequence and correlation
For a comprehensive energy system containing multiple renewable energy sources and multiple energy loads, the application of the method for generating the scene by considering the time sequence, the autocorrelation and the cross correlation mainly comprises the following parts:
1) and carrying out scene data normalization processing according to the historical data. Based on the data, determining the scene clustering number by adopting a Davies-Bouldin index, and clustering the scenes by utilizing a K-means method;
2) establishing a probability distribution model of the multi-energy load, and fitting distribution parameters by adopting a maximum likelihood estimation method according to the established probability distribution model;
3) analyzing time sequence autocorrelation characteristics under different scenes, and generating a scene with autocorrelation by adopting an LHS-CD method;
4) extracting scenes from the scene clusters with autocorrelation under different source load categories, combining to generate a scene set, and calculating the cross correlation of the scene set. And (3) screening out a cross-correlation scene set which is relatively close to the cross-correlation of historical data on an MATLAB simulation platform by adopting a particle swarm optimization to obtain a final multi-energy load scene set with time sequence and correlation.
The scene generation is analyzed in the following example, which is described in detail below:
In order to research the effectiveness of a scene production method, the scene generation method is implemented under a certain example comprehensive energy scene, renewable energy and multi-energy loads in a region are selected as objects to perform scene generation, 3 conditions are set for comparison, and the condition 1 is a historical data result and is indicated by HD; case 2 is that a scene result is generated by using a Monte Carlo sampling method, and MC is used for indicating; case 3 is using the autoregressive moving average model method, denoted by ARMA; case 4 is a scenario generation method using the proposed multi-energy charge-timing sequence, correlation, referred to as MDCS.
The scene clustering results and distribution parameters generated in the first two steps of the MDCS method are shown in fig. 3. The result pairs under different scene generation methods are shown in fig. 4 and 5. The results of fig. 4 show that the proposed scene generation Method (MDCS) considering multi-energy charge-time sequence, correlation is closer to the Historical Data (HD) than the monte carlo sampling Method (MC) and the autoregressive moving average model method (ARMA). FIG. 5 shows that the method of the present invention is indeed superior to the general method under different evaluation indexes.
The scene error comparison result of the scene generation method and the historical data is shown in table 1. As can be seen from the table, the proposed method is closer to the historical data scenario in terms of both Probability Distribution Function (PDF) and Cumulative Distribution Function (CDF), and has smaller error.
TABLE 1 Scenario generation formula considering multi-energy charge-time sequence, correlation and historical data error comparison
Figure BDA0002519778190000081
Those skilled in the art will appreciate that the drawings are only schematic illustrations of preferred embodiments, and the above-described embodiments of the present invention are merely provided for description and do not represent the merits of the embodiments.
The present invention is not limited to the above-described embodiments. The foregoing description of the specific embodiments is intended to describe and illustrate the technical solutions of the present invention, and the above specific embodiments are merely illustrative and not restrictive. Those skilled in the art can make many changes and modifications to the invention without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (7)

1. A scene generation method considering multi-energy charge-time sequence and correlation is characterized by comprising the following steps:
1) carrying out scene data normalization processing according to historical data; based on the data after normalization processing, determining the scene clustering number by adopting a Davies-Bouldin index, and carrying out scene clustering by utilizing a K-means method;
2) establishing a probability distribution model of the multi-energy load, and fitting distribution parameters by adopting a maximum likelihood estimation method according to the established probability distribution model;
3) Analyzing time sequence autocorrelation characteristics under different scenes, and generating a scene with autocorrelation by adopting an LHS-CD method; extracting scenes from scene clusters with autocorrelation under different source load categories, combining to generate a scene set, and calculating the cross correlation of the scene set;
4) and (3) screening out a cross-correlation scene set which is close to the cross-correlation with the historical data on an MATLAB simulation platform by adopting a particle swarm optimization to obtain a final multi-energy load scene set with time sequence and correlation.
2. The scene generation method considering the multi-energy charge-timing sequence and the correlation according to claim 1, wherein the Davies-Bouldin index is specifically:
Figure FDA0002519778180000011
in the formula
Figure FDA0002519778180000012
Representing the distance from the ith cluster sample to the cluster center;
Figure FDA0002519778180000013
representing the distance from the jth cluster sample to the cluster center; NUMjClustering number of the jth cluster samples; w is aiRepresenting the ith cluster centroid; w is ajRepresenting the jth cluster centroid; i Wi-wj||2Representing the distance between the centroids of the ith cluster and the jth cluster; the Davies-Bouldin index reflects the mean value of the maximum similarity in each class, and the smaller the value, the better the clustering effect.
3. The scene generation method considering the multi-energy charge-time sequence and correlation according to claim 1, wherein the formula involved in analyzing the time sequence autocorrelation characteristics under different scenes is specifically as follows:
Figure FDA0002519778180000014
Figure FDA0002519778180000015
Wherein s isiIs the ith scene value;
Figure FDA0002519778180000016
a clustering center for the kth seed charge; r isi,i+gIs the element of the ith row and the ith + g column of the autocorrelation matrix; τ is the number of time sequences; rhoautoIs an autocorrelation matrix.
4. The scene generation method considering the timing sequence and the correlation of the multi-energy loads according to claim 1, wherein the probability distribution model of the multi-energy loads is specifically:
Figure FDA0002519778180000021
in the formula
Figure FDA0002519778180000022
A mathematical expectation representing k energy loads at time t;
Figure FDA0002519778180000023
the variance of k energy loads at t moment is represented;
Figure FDA0002519778180000024
the load value of the kth load at the time t is shown;
Figure FDA0002519778180000025
is the upper limit of the load value;
Figure FDA0002519778180000026
is the lower limit of the load value;
Figure FDA0002519778180000027
is the probability density function of the k load at the time t.
5. The scene generation method considering the timing sequence and the correlation of the multi-energy loads according to claim 1 or 4, wherein the source in the multi-energy loads is renewable energy, including wind power and photovoltaic power; the probability distribution model of wind power and photovoltaic specifically comprises the following steps:
Figure FDA0002519778180000028
wherein v represents wind speed; k represents a shape parameter of the Weibull distribution; c represents a scale parameter of the Weibull distribution; (v) a probability density function representing wind speed;
Figure FDA0002519778180000029
wherein S represents the intensity of light; srRepresenting a rated illumination intensity; α, β represent distribution parameters of the beta distribution; f (S) represents a probability density function of the illumination intensity.
6. The scene generation method considering the multi-energy charge-time sequence and correlation according to claim 1, wherein the LHS-CD method is specifically:
assume m random variables X with correlation1,X2,…,Xi,…,XmAnd XiHas a probability distribution of Yi＝F(Xi)；
1) And (3) sampling matrix X extraction: n sample points are extracted for each random variable, at the moment, an Nxm-order sampling matrix is extracted, and in order to ensure that the extracted samples uniformly cover the whole adopted interval, the random variable X is usediThe probability distribution function is divided into m subintervals, the probability corresponding to each interval is equal, and a sampling value is extracted from each interval; to avoid direct processing of the probability distribution function, Y is addedi＝F(Xi) The longitudinal axis of the method is divided into N parts averagely, N is the average number of the LHS method, the length of each interval is 1/N, one sample point is randomly taken from N intervals which are not intersected with each other, and then Y is used for obtaining the average number of the LHS methodiThe inverse function of (A) to (B) to obtain a random variable XiOf the sampled value, thus the random variable XiThe nth sample value of (d) is:
Figure FDA0002519778180000031
forming N sampling values of each random variable into a row of a matrix to obtain an Nxm-order matrix;
2) randomly generating a sequence matrix L, and calculating a Spearman correlation coefficient matrix rho of the L by adopting a time sequence autocorrelation characteristic formulaLResolving and eliminating the correlation generated by random arrangement by adopting Cholesky; target correlation coefficient matrix ρ for historical data objCholesky decomposition is carried out to enable random arrangement of the sequence matrix L to generate corresponding correlation; updating the initial sample matrix X to obtain a new sample matrix XuMake the arrangement order of each row of elements and GuThe element sequence of corresponding row in the sequence table is the same, and the obtained XuHaving a target relevance;
ρL＝QQT
G＝Q-1L
ρobj＝PPT
Gu＝PG＝PQ-1L
where rhoLA correlation matrix which is an order matrix L; q is a decomposed diagonal matrix; qTA transposed matrix that is Q; q-1An inverse matrix of Q; g is a transition matrix; rhoobjIs a target correlation matrix; p is a decomposed diagonal matrix; guSamples with a specified target correlation.
7. The scene generation method considering the multi-energy charge-time sequence and the correlation according to claim 1, wherein the cross-correlation analysis method specifically comprises:
for two time sequential scenarios
Figure FDA0002519778180000032
And
Figure FDA0002519778180000033
the cross-correlation matrix is:
Figure FDA0002519778180000034
Figure FDA0002519778180000035
wherein M is the number of source charge variables; si,sjTime sequence scene values of the ith and jth moments;
Figure FDA0002519778180000041
a clustering center for the kth seed charge;
Figure FDA0002519778180000042
a clustering center of the u-th seed and load; ri,jIs the cross-correlation coefficient; rhocrossIs a cross correlation matrix.
CN202010487888.1A 2020-06-02 2020-06-02 Scene generation method considering multi-energy charge time sequence and correlation Active CN111859283B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010487888.1A CN111859283B (en) 2020-06-02 2020-06-02 Scene generation method considering multi-energy charge time sequence and correlation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010487888.1A CN111859283B (en) 2020-06-02 2020-06-02 Scene generation method considering multi-energy charge time sequence and correlation

Publications (2)

Publication Number Publication Date
CN111859283A true CN111859283A (en) 2020-10-30
CN111859283B CN111859283B (en) 2024-04-02

Family

ID=72985392

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010487888.1A Active CN111859283B (en) 2020-06-02 2020-06-02 Scene generation method considering multi-energy charge time sequence and correlation

Country Status (1)

Country Link
CN (1) CN111859283B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113190973A (en) * 2021-04-09 2021-07-30 国电南瑞科技股份有限公司 Bidirectional optimization method, device, equipment and storage medium for wind, light and load multi-stage typical scene
CN113553709A (en) * 2021-07-22 2021-10-26 杭州电子科技大学 Scene modeling method suitable for planning operation of comprehensive energy system
CN116227751A (en) * 2023-05-09 2023-06-06 国网吉林省电力有限公司经济技术研究院 Optimal configuration method and device for power distribution network

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080183786A1 (en) * 2007-01-30 2008-07-31 International Business Machines Corporation Systems and methods for distribution-transition estimation of key performance indicator
JP2012023816A (en) * 2010-07-12 2012-02-02 Toshiba Corp Information processing equipment, and program thereof
CN109508823A (en) * 2018-11-06 2019-03-22 西安理工大学 A kind of Distributed Generation in Distribution System planing method of method based on scene analysis
CN110380444A (en) * 2019-06-26 2019-10-25 东南大学 Distributing wind-powered electricity generation orderly accesses the method for planning capacity of power grid under a kind of more scenes based on structure changes Copula

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080183786A1 (en) * 2007-01-30 2008-07-31 International Business Machines Corporation Systems and methods for distribution-transition estimation of key performance indicator
JP2012023816A (en) * 2010-07-12 2012-02-02 Toshiba Corp Information processing equipment, and program thereof
CN109508823A (en) * 2018-11-06 2019-03-22 西安理工大学 A kind of Distributed Generation in Distribution System planing method of method based on scene analysis
CN110380444A (en) * 2019-06-26 2019-10-25 东南大学 Distributing wind-powered electricity generation orderly accesses the method for planning capacity of power grid under a kind of more scenes based on structure changes Copula

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
杨卫红;赖清平;兰宇;王丹;胡庆娥;王旭阳;刘艳茹;: "基于调节潜力指标的用户用电行为聚类分析算法研究", 电力建设, no. 06 *
江雪辰;朱俊澎;袁越;王跃峰;黄阮明;: "基于新型场景划分与考虑时序相关性的光伏出力时间序列模拟方法", 电力建设, no. 10 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113190973A (en) * 2021-04-09 2021-07-30 国电南瑞科技股份有限公司 Bidirectional optimization method, device, equipment and storage medium for wind, light and load multi-stage typical scene
CN113553709A (en) * 2021-07-22 2021-10-26 杭州电子科技大学 Scene modeling method suitable for planning operation of comprehensive energy system
CN116227751A (en) * 2023-05-09 2023-06-06 国网吉林省电力有限公司经济技术研究院 Optimal configuration method and device for power distribution network
CN116227751B (en) * 2023-05-09 2023-07-07 国网吉林省电力有限公司经济技术研究院 Optimal configuration method and device for power distribution network

Also Published As

Publication number Publication date
CN111859283B (en) 2024-04-02

Similar Documents

Publication Publication Date Title
Zhao et al. Correlation-constrained and sparsity-controlled vector autoregressive model for spatio-temporal wind power forecasting
Liu et al. Short-term wind speed forecasting based on spectral clustering and optimised echo state networks
CN111859283B (en) Scene generation method considering multi-energy charge time sequence and correlation
CN111401755B (en) Multi-new-energy output scene generation method, device and system based on Markov chain
CN108717579B (en) Short-term wind power interval prediction method
CN112149264A (en) Active power distribution network planning method based on improved Harris eagle optimization algorithm
CN112100911B (en) Solar radiation prediction method based on depth BILSTM
CN107679687A (en) A kind of photovoltaic output modeling method and Generation System Reliability appraisal procedure
CN113065278B (en) Frequent pattern mining-based prediction method for wind power small-occurrence event statistical characteristic model in rich period
CN110516944A (en) A kind of power distribution network multistage typical case's Run-time scenario generation method
Gu et al. Short-term wind power forecasting and uncertainty analysis based on FCM–WOA–ELM–GMM
Hu et al. Short-term photovoltaic power prediction based on similar days and improved SOA-DBN model
CN104951654A (en) Method for evaluating reliability of large-scale wind power plant based on control variable sampling
Guo et al. A meteorological information mining-based wind speed model for adequacy assessment of power systems with wind power
CN112418504B (en) Wind speed prediction method based on mixed variable selection optimization deep belief network
Putra et al. Modeling of high uncertainty photovoltaic generation in quasi dynamic power flow on distribution systems: A case study in Java Island, Indonesia
CN112508278A (en) Multi-connected system load prediction method based on evidence regression multi-model
CN108694475B (en) Short-time-scale photovoltaic cell power generation capacity prediction method based on hybrid model
Zhen et al. Simultaneous prediction for multiple source–loads based sliding time window and convolutional neural network
Sun et al. Load-forecasting method for IES based on LSTM and dynamic similar days with multi-features
CN113159437B (en) Method for predicting short-term photovoltaic power generation output power
Prashanthi et al. A comparative study of the performance of machine learning based load forecasting methods
Wang et al. An alternative method for estimating wind-power capacity credit based on reliability evaluation using intelligent search
Guan et al. Multiple wind power time series modeling method considering correlation
CN113222216A (en) Method, device and system for predicting cooling, heating and power loads

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