CN110336332B - Interval power flow typical scene construction method based on output curve aggregation - Google Patents

Interval power flow typical scene construction method based on output curve aggregation Download PDF

Info

Publication number
CN110336332B
CN110336332B CN201910693437.0A CN201910693437A CN110336332B CN 110336332 B CN110336332 B CN 110336332B CN 201910693437 A CN201910693437 A CN 201910693437A CN 110336332 B CN110336332 B CN 110336332B
Authority
CN
China
Prior art keywords
curve
clustering
output
curves
power
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
CN201910693437.0A
Other languages
Chinese (zh)
Other versions
CN110336332A (en
Inventor
刘丽军
笪超
罗宁
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.)
Fuzhou University
Original Assignee
Fuzhou 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 Fuzhou University filed Critical Fuzhou University
Priority to CN201910693437.0A priority Critical patent/CN110336332B/en
Publication of CN110336332A publication Critical patent/CN110336332A/en
Application granted granted Critical
Publication of CN110336332B publication Critical patent/CN110336332B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H02J3/382—
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention relates to an interval power flow typical scene construction method based on output curve aggregation, which comprises the following steps of S1: substituting the wind speed and illumination intensity data into a wind-solar output formula to obtain an RDG (remote data generator) solar output curve sample; step S2: randomly selecting a curve from the samples as a first curve in the initial center; step S3: constructing an initial center; step S4: calculating Euclidean distances from the residual curves to K clustering centers, dividing the residual curves into corresponding clustering categories, and generating a K-category curve set; step S5: updating K clustering center curves; step S6: judging whether a convergence condition is reached; if yes, executing step S7, otherwise returning to step S4; step S7: adopting different K value clusters, calculating profile parameters of corresponding results, and selecting an optimal cluster K value; step S8: generating an interval power flow typical scene; step S9: and calculating the interval power flow according to the 24 XK groups of hour output intervals. The method can avoid the problem of section ultra-width of section power flow in the whole sample space sampling calculation section.

Description

Interval power flow typical scene construction method based on output curve aggregation
Technical Field
The invention relates to the field of new energy, in particular to a method for constructing an interval tide typical scene based on output curve aggregation.
Background
To address the increasingly severe global energy crisis and environmental issues, large-scale, high permeability Distributed Renewable energy (RDG) grid integration has become a necessary trend. Distributed energy such as wind and light is influenced by factors such as geographical position, natural environment, weather conditions and the like, and the output characteristic is often strong in randomness and volatility.
The existing deterministic load flow calculation method is difficult to consider uncertainty, and the result of the method is often different from the follow-up actual operation condition.
The method is an effective uncertainty optimization planning means based on interval load flow calculation and establishment of an interval optimization model, but in order to obtain a global optimal solution, the interval load flow has to be calculated on a full-state space, so that the conservation of the interval load flow calculation is greatly expanded, and the calculation result loses the engineering application value.
The most common methods currently used to consider uncertainty are three main types: the method comprises the steps of performing multi-objective optimization through probability power flow embedded Monte Carlo simulation based on a new energy output probability model, performing interval optimization through calculation of interval power flow based on a new energy output interval, and converting an uncertainty problem into a certainty problem based on various typical operation scenes. The first method needs to obtain a probability model of power output to calculate the probability load flow, has the problems of high data acquisition difficulty, inaccurate probability parameters, large calculated amount and the like, and limits the application of the method. The second method is to calculate the load flow in the form of interval number, and the interval load flow can cover all uncertain conditions, but has a certain conservative problem. The third method constructs a typical scene, the uncertainty load flow calculation is decomposed into a plurality of certainty load flow calculations, and how to select a representative scene to ensure that all uncertainty conditions can be covered is a long-standing difficulty of the method. The current method is difficult to simultaneously ensure the computation complexity and the computation precision of the model when considering the uncertainty problem to compute the system load flow.
Disclosure of Invention
In view of this, the present invention provides a method for constructing a typical interval power flow scene based on aggregation of output curves, which can avoid the problem of interval superwidth of interval power flow in full-sample space sampling calculation.
The invention is realized by adopting the following scheme: an interval power flow typical scene construction method based on output curve aggregation comprises the following steps:
step S1: enabling a period to have L days, providing wind speed and illumination intensity data of the period, substituting the data into a wind-solar output formula for calculation, and obtaining L wind-solar output curves in the period;
step S2: the L wind and light solar output curves are expressed as NiThe method comprises the following steps that (1), psi (2), psi (L) and each curve is an n-dimensional vector and contains 24 moments of wind power and photovoltaic output data; taking L wind and light solar output curves as an original sample set NiSetting the number of clusters K, and randomly selecting NiOne curve is selected
Figure BDA0002148075820000021
As an initial center T(0)The first curve of (1);
step S3: calculate each curve and in the remaining samples according to equation (1)
Figure BDA0002148075820000022
European distance of
Figure BDA0002148075820000023
Selecting the curve with the maximum d as a second initial central curve
Figure BDA0002148075820000024
Initial center at this time
Figure BDA0002148075820000031
Figure BDA0002148075820000032
In the formula
Figure BDA0002148075820000033
Indicating the distance between the curves,. psii(k) Is the k-th dimension data of curve i,
Figure BDA0002148075820000034
representing the jth curve of the clustering center;
for the remaining sample set Mi(Mi∈Ni/T(0)) Calculate each curve to T(0)Taking the curve with the maximum sum of the distances as the next initial central curve; executing K times of maximum distance calculation to obtain K initial clustering center curves
Figure BDA0002148075820000035
Step S4: computing an original sample set NiRemoving the residual curve of the cluster center curve to K cluster center curves
Figure BDA0002148075820000036
The Euclidean distance of the curve is obtained by classifying the curves closest to the clustering center into a class of scenes, so that K curve sets are obtained, and each curve set is called a scene;
step S5: calculating the average value of data of each output curve in K scenes at the same moment to obtain a new clustering center curve, and updating K clustering center curves;
step S6: when the K clustering center curves are not changed any more, the clustering is considered to be convergent, the L curve samples are divided into K curve sets, and each curve set comprises a plurality of curves; judging whether a convergence condition is reached; if yes, executing step S7, otherwise returning to step S4;
step S7, the clustering sample is composed of L wind and light solar output curves, the clustering number K takes the value as the range
Figure BDA0002148075820000037
The integers in the step S2 to the step S6 are executed by adopting different clustering numbers K, the total contour parameters of the results after clustering of different K values are calculated, the size of the total contour parameters is compared to select the optimal K value, and the K scenes obtained by executing the steps S2 to S6 according to the K value are the clustering results;
step S8: constructing 24 XK groups of hourly output intervals according to RDG (remote data generator) daily output curve data in each scene to generate an interval power flow typical scene;
step S9: and calculating the section power flow according to the 24 XK groups of hour output intervals obtained in the step S8.
Further, the specific content of step S1 is:
providing wind speed and light intensity data v for a period of timei,t、hi,tWherein i is 1,2,3, …, L is the number of days in the period, 0 < L < 3650, t represents 24 hours; the data is substituted into equations (2) and (3) to be calculated.
Figure BDA0002148075820000041
In the formula Pwi,tOutput power, v, for wind power generation at the ith time of dayi,tRepresents the wind speed at the time t on the ith day, vs、vrAnd v0Respectively representing rated wind speed, cut-in wind speed and cut-out wind speed;
Figure BDA0002148075820000042
in the formula Pvi,tOutput power h of photovoltaic power generation at the ith momenti,tRepresents the actual illumination intensity at the t time of the ith day, hrIndicating the nominal light intensity, PVRRated output power;
calculating to obtain the time sequence data P of the wind power output of the ith dayW(i)＝(Pwi,1,Pwi,2,…,Pwi,24) And photovoltaic solar output time sequence data PV(i)＝(Pvi,1,Pvi,2,…,Pvi,24) The power curve psi (i) is combined as Pwi,1,Pwi,2,…,Pwi,24,Pvi,1,Pvi,2,…,Pvi,24]The sunrise force curves { ψ (1), ψ (2),.., ψ (L) } for L days are obtained.
Further, the step S7 specifically includes the following steps:
step S71, the clustering sample is composed of L wind and light solar output curves, the clustering number K takes the value as the range
Figure BDA0002148075820000053
Performing steps S2 to S6 on the sample curve by adopting different K values to obtain corresponding K clustering results;
step S72, for each type of K value clustered clustering result, passing the contour parameter S (psi)i) Quantitative clustering effect, contour parameter S (psi)i) The value range is [ -1,1 [ ]]；
Figure BDA0002148075820000051
In the formula, a (psi)i) Representing the Euclidean distance mean value of the curve and other curves in the same type of scene as the clustering degree; b (psi)i) Representing the Euclidean distance mean value of the curve and all curves in other scenes as the clustering separation degree;
total profile parameter StIs defined as:
Figure BDA0002148075820000052
step S73, comparing the total contour parameters S obtained by calculating the formula (4) and the formula (5) under different K value clustering resultstAnd (4) the clustering result with the maximum total contour parameter has the optimal clustering effect, the corresponding K value is the optimal clustering K value, and the K scenes obtained in the steps from S2 to S6 are the clustering result according to the K value.
Further, the step S8 specifically includes the following steps:
let s RDG sunrise force curves in the k scene, as shown in formula (6), and the output sequence P at the t moment in the scenew,t＝{Pwr,t,Pwo,t,...,Pws,t}; will sequence Pw,tThe output data is rearranged into a new sequence Y ═ Y from small to large1,y2,...,ysProcessing the sequence Y according to formula (7) to obtain YiBy the value z thereiniInstead of obtaining the sequence Z ═ Z1,z2,...,zs}; obtaining a new output interval [ minZ, maxZ ] based on the data in the sequence Z]；
Figure BDA0002148075820000061
Figure BDA0002148075820000062
Processing the 24-hour output data of the scene to obtain 24 groups of hour RDG output intervals
Figure BDA0002148075820000063
In the formula [ Prdg,i]Representing the output interval of the RDG in the scene at the ith moment of the scene; and sequentially carrying out interval construction on the K scenes to obtain 24 XK groups of hour output intervals.
Further, the step S9 specifically includes the following steps:
step S91, calculating the power of the node:
Figure BDA0002148075820000064
in the formula SiRepresenting the node computation power, SLiInjecting power, y, into the nodei0Admittance for the terrestrial branch;
step S92, starting from the end of the distribution network, gradually advancing, and solving the power S of each branch of the distribution network by the node voltageijDistributing;
Figure BDA0002148075820000065
Figure BDA0002148075820000066
in the formula [ Sij]For branch power, Δ SijFor line power loss, Pj、QjFor node active and reactive power, ZijIs line impedance, SijIs the branch power;
step S93, calculating the node voltage interval from the branch power
Figure BDA0002148075820000071
Figure BDA0002148075820000072
Figure BDA0002148075820000073
In the formula
Figure BDA0002148075820000074
Respectively representing the longitudinal and transverse components of the branch pressure drop;
Figure BDA0002148075820000075
Figure BDA0002148075820000076
step S94, convergence criterion: the iteration termination criterion is the voltage interval [ U ] of each nodei]The deviation of the upper and lower limits relative to the value of the last iteration is smaller than an allowable value as shown in the formula (15);
Figure BDA0002148075820000077
in the formula
Figure BDA0002148075820000078
Respectively representing the upper limit and the lower limit of the voltage amplitude of the node j in the kth time, wherein epsilon is a deviation allowable value;
after convergence, the voltage interval [ U ] of each node is obtainedi]Namely, the interval power flow calculation result.
Compared with the prior art, the invention has the following beneficial effects: clustering is carried out by taking the RDG output curve as a sample, an hour output interval is constructed, an interval trend typical scene is further obtained, the annual operation characteristics of a planning area can be reflected, uncertainty factors are fully considered in the planning process, and the problem of conservative expansion caused by the trend in a full-state space calculation interval is avoided.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Detailed Description
The invention is further explained below with reference to the drawings and the embodiments.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
As shown in fig. 1, the present embodiment provides a method for constructing an interval power flow typical scene based on an output curve aggregation, including the following steps:
step S1: enabling a period to have L days, providing wind speed and illumination intensity data of the period, substituting the data into a wind-solar output formula for calculation, and obtaining L all wind-solar output curves and L all light-solar output curves in the period;
step S2: the L wind and light solar output curves are expressed as NiThe method comprises the following steps that (1), psi (2), psi (L) and each curve is an n-dimensional vector and contains 24 moments of wind power and photovoltaic output data; taking L wind and light solar output curve curves as an original sample set NiSetting the number of clusters K, and randomly selecting NiSelecting a curve
Figure BDA0002148075820000085
As an initial center T(0)The first curve of (1);
step S3: according to formula (1) inCalculating the sum of each curve in the rest samples
Figure BDA0002148075820000081
European distance of
Figure BDA0002148075820000082
Selecting the curve with the maximum d as a second initial central curve
Figure BDA0002148075820000083
Initial center at this time
Figure BDA0002148075820000084
Figure BDA0002148075820000091
In the formula
Figure BDA0002148075820000092
Indicating the distance between the curves,. psii(k) Is the k-th dimension data of curve i,
Figure BDA0002148075820000093
representing the jth curve of the clustering center;
for the remaining sample set Mi(Mi∈Ni/T(0)) Calculate each curve to T(0)Taking the curve with the maximum sum of the distances as the next initial central curve;
k times of maximum distance calculation is executed according to the method, and then K initial clustering center curves are obtained
Figure BDA0002148075820000094
Step S4: computing an original sample set NiThe Euclidean distance from the rest curves except the clustering center curve to the K clustering center curves is classified into a class of scenes, and the K curves are obtained by classifying the curves and the curves closest to the clustering centerEach curve set is called a scene;
step S5: calculating the average value of data of each output curve in K scenes at the same moment to obtain a new clustering center curve, and updating K clustering center curves;
step S6: when the K clustering center curves are not changed any more, the clustering is considered to be convergent, the L curve samples are divided into K curve sets, and each curve set comprises a plurality of curves; judging whether a convergence condition is reached; if yes, executing step S7, otherwise returning to step S4;
step S7: the clustering sample consists of L wind and light output curves, and the clustering number K takes the value as the range
Figure BDA0002148075820000095
And (4) adopting different clustering numbers K to execute the steps S2 to S6, calculating total contour parameters of the results after clustering with different K values, comparing the sizes of the total contour parameters to select an optimal K value, and executing the K scenes obtained in the steps S2 to S6 according to the K value to obtain clustering results.
Step S8: constructing 24 XK groups of hourly output intervals according to RDG (remote data generator) daily output curve data in each scene to generate an interval power flow typical scene;
step S9: and calculating the section power flow according to the 24 XK groups of hour output intervals obtained in the step S8.
In this embodiment, the specific content of step S1 is:
providing wind speed and light intensity data v of a periodi,t、hi,tWhere i is 1,2,3, …, L being the number of days in the cycle, 0 < L < 3650, t representing 24 hours. The data is substituted into equations (2) and (3) to be calculated.
Figure BDA0002148075820000101
In the formula Pwi,tOutput power, v, for wind power generation at the ith time of dayi,tRepresents the wind speed at the time t on the ith day, vs、vrAnd v0The rated wind speed, cut-in wind speed and cut-out wind speed are respectively represented.
Figure BDA0002148075820000102
In the formula Pvi,tOutput power h of photovoltaic power generation at the ith momenti,tRepresents the actual illumination intensity at the t time of the ith day, hrIndicating the nominal light intensity, PVRIs the rated output power.
Calculating to obtain the time sequence data P of the wind power output of the ith dayW(i)＝(Pwi,1,Pwi,2,…,Pwi,24) And photovoltaic solar output time sequence data PV(i)＝(Pvi,1,Pvi,2,…,Pvi,24) The power curve psi (i) is combined as Pwi,1,Pwi,2,…,Pwi,24,Pvi,1,Pvi,2,…,Pvi,24]The sunrise force curves { ψ (1), ψ (2),. -, ψ (L) } for L days in the period are obtained.
In this embodiment, the step S7 specifically includes the following steps:
step S71, the clustering sample is composed of L wind and light solar output curves, the clustering number K takes the value as the range
Figure BDA0002148075820000111
Performing steps S2 to S6 on the sample curve by adopting different K values to obtain corresponding K clustering results;
step S72, for each type of K value clustered clustering result, passing the contour parameter S (psi)i) Quantitative clustering effect, contour parameter S (psi)i) The value range is [ -1,1 [ ]]。
Figure BDA0002148075820000112
In the formula, a (psi)i) Representing the Euclidean distance mean value of the curve and other curves in the same type of scene as the clustering degree; b (psi)i) To cluster degrees of separation, tableShowing the mean of the euclidean distances of the curve from all curves in other scenes.
Total profile parameter StIs defined as:
Figure BDA0002148075820000113
step S73, comparing the total contour parameters S obtained by calculating the formula (4) and the formula (5) under different K value clustering resultstAnd (4) the clustering result with the maximum total contour parameter has the optimal clustering effect, the corresponding K value is the optimal clustering K value, and the K scenes obtained in the steps from S2 to S6 are the clustering result according to the K value.
In this embodiment, the step S8 specifically includes the following steps:
let s RDG sunrise force curves in the k scene, as shown in formula (6), and the output sequence P at the t moment in the scenew,t＝{Pwr,t,Pwo,t,...,Pws,t}; will sequence Pw,tThe output data is rearranged into a new sequence Y ═ Y from small to large1,y2,...,ysProcessing the sequence Y according to formula (7) to obtain YiBy the value z thereiniInstead of obtaining the sequence Z ═ Z1,z2,...,zs}; obtaining a new output interval [ minZ, maxZ ] based on the data in the sequence Z]。
The 24-hour output data of the scene is processed by the method to obtain 24 groups of hour wind and light output intervals
Figure BDA0002148075820000121
In the formula [ Pw,1]、[Pv,1]Respectively representing the output intervals of wind power and photovoltaic at the 1 st moment of the scene. Constructing intervals of K scenes according to the method to obtain 24 multiplied by K groups of hour output intervals;
Figure BDA0002148075820000122
Figure BDA0002148075820000123
in this embodiment, the step S9 specifically includes the following steps:
and after K typical scenes are obtained through clustering and 24 XK groups of hourly output intervals are determined, the power flow of the distribution network interval is respectively calculated. Knowing the voltage of the starting end and the power of the tail end of the power distribution network, firstly, assuming that the voltage of the power distribution network is rated voltage, and gradually carrying out forward push from the tail end in a forward push process to calculate the power distribution of branches of the whole network; the backward substitution process is gradually pushed back from the starting end, and the voltage of each node is calculated; and continuously repeating the forward pushing step and the backward replacing step until convergence.
Number of intervals
Figure BDA0002148075820000124
WhereinxAnd
Figure BDA0002148075820000125
are respectively [ X ]]The lower limit and the upper limit of (2) are given by the symbols [, ]]To indicate the number of intervals; the step S9 is performed by using interval numbers, and specifically includes the following steps:
and step S91, considering the influence on the ground admittance branch, calculating the power by the calculation node:
Figure BDA0002148075820000126
in the formula SiRepresenting the node computation power, SLiInjecting power, y, into the nodei0Admittance for the terrestrial branch;
and step S92, starting from the end of the power distribution network, gradually advancing, and solving the power distribution of each branch of the power distribution network by the node voltage.
Figure BDA0002148075820000131
Figure BDA0002148075820000132
In the formula [ Sij]For branch power, Δ SijFor line power loss, Pj、QjFor node active and reactive power, ZijIs line impedance, SijIs the branch power;
step S93, calculating the node voltage interval from the branch power
Figure BDA0002148075820000133
Figure BDA0002148075820000134
Figure BDA0002148075820000135
In the formula
Figure BDA0002148075820000136
Respectively representing the longitudinal and transverse components of the branch pressure drop;
Figure BDA0002148075820000137
Figure BDA0002148075820000138
step S94, convergence criterion: the iteration termination criterion is the voltage interval [ U ] of each nodei]The deviation of the upper limit and the lower limit relative to the value of the last iteration is smaller than an allowable value;
Figure BDA0002148075820000139
in the formula
Figure BDA00021480758200001310
Respectively representing the upper limit and the lower limit of the voltage amplitude of the node j in the kth time, wherein epsilon is a deviation allowable value;
after convergence, the voltage interval [ U ] of each node is obtainedi]Namely, the interval power flow calculation result.
In the embodiment, the RDG output curve is used as a sample for clustering, an hour output interval is constructed, an interval trend typical scene is further obtained, the annual operation characteristics of a planning area can be reflected, uncertainty factors are fully considered in the planning process, and the problem of conservative expansion caused by the trend in a full-state space calculation interval is solved. The technical effects of this embodiment are shown in table 1 in conjunction with specific embodiments.
TABLE 1 comparison of interval load flow calculation results of different methods
Figure BDA0002148075820000141
In this embodiment (1), the euclidean distance is used as the similarity evaluation criterion, and the RDG output curve is regarded as a multidimensional vector for clustering division;
(2) removing noise points in the curves based on a median method aiming at the clustered curve set, and providing an hour output interval construction method to generate an interval tide typical scene;
(3) the interval load flow is calculated based on the hour output interval in a typical scene, and technical support is provided for the power distribution network to process the uncertainty problem.
The above description is only a preferred embodiment of the present invention, and all equivalent changes and modifications made in accordance with the claims of the present invention should be covered by the present invention.

Claims (3)

1. An interval power flow typical scene construction method based on output curve aggregation is characterized by comprising the following steps: the method comprises the following steps:
step S1: enabling a period to have L days, providing wind speed and illumination intensity data of the period, substituting the data into a wind-solar output formula for calculation, and obtaining L wind-solar output curves in the period;
step S2: the L wind and light solar output curves are expressed as NiThe method comprises the following steps that (1), psi (2), psi (L) and each curve is an n-dimensional vector and contains 24 moments of wind power and photovoltaic output data; taking L wind and light solar output curves as an original sample set NiSetting the number of clusters K, and randomly selecting NiOne curve is selected
Figure FDA0002837860680000011
As an initial center T(0)The first curve of (1);
step S3: calculate each curve and in the remaining samples according to equation (1)
Figure FDA0002837860680000012
European distance of
Figure FDA0002837860680000013
Selecting the curve with the maximum d as a second initial central curve
Figure FDA0002837860680000014
Initial center at this time
Figure FDA0002837860680000015
Figure FDA0002837860680000016
In the formula
Figure FDA0002837860680000017
Indicating the distance between the curves,. psii(k) Is the k-th dimension data of curve i,
Figure FDA0002837860680000018
representing the jth curve of the clustering center;
for the remaining sample set Mi(Mi∈Ni/T(0)) Calculate each curve to T(0)Taking the curve with the maximum sum of the distances as the next initial central curve; executing K times of maximum distance calculation to obtain K initial clustering center curves
Figure FDA0002837860680000019
Step S4: computing an original sample set NiRemoving the residual curve of the cluster center curve to K cluster center curves
Figure FDA00028378606800000110
The Euclidean distance of the curve is obtained by classifying the curves closest to the clustering center into a class of scenes, so that K curve sets are obtained, and each curve set is called a scene;
step S5: calculating the average value of data of each output curve in K scenes at the same moment to obtain a new clustering center curve, and updating K clustering center curves;
step S6: when the K clustering center curves are not changed any more, the clustering is considered to be convergent, the L curve samples are divided into K curve sets, and each curve set comprises a plurality of curves; judging whether a convergence condition is reached; if yes, executing step S7, otherwise returning to step S4;
step S7, the clustering sample is composed of L wind and light solar output curves, the clustering number K takes the value as the range
Figure FDA0002837860680000022
The integers in the step S2 to the step S6 are executed by adopting different clustering numbers K, the total contour parameters of the results after clustering of different K values are calculated, the size of the total contour parameters is compared to select the optimal K value, and the K scenes obtained by executing the steps S2 to S6 according to the K value are the clustering results;
step S8: constructing 24 XK groups of hourly output intervals according to RDG (remote data generator) daily output curve data in each scene to generate an interval power flow typical scene;
step S9: calculating an interval power flow according to the 24 xK groups of hour output intervals obtained in the step S8;
wherein, the step S7 specifically includes the following contents:
step S71, the clustering sample is composed of L wind and light solar output curves, the clustering number K takes the value as the range
Figure FDA0002837860680000023
Performing steps S2 to S6 on the sample curve by adopting different K values to obtain corresponding K clustering results;
step S72, for each type of K value clustered clustering result, passing the contour parameter S (psi)i) Quantitative clustering effect, contour parameter S (psi)i) The value range is [ -1,1 [ ]]；
Figure FDA0002837860680000021
In the formula, a (psi)i) Representing the Euclidean distance mean value of the curve and other curves in the same type of scene as the clustering degree; b (psi)i) Representing the Euclidean distance mean value of the curve and all curves in other scenes as the clustering separation degree;
total profile parameter StIs defined as:
Figure FDA0002837860680000031
step S73, comparing the total contour parameters S obtained by calculating the formula (4) and the formula (5) under different K value clustering resultstThe clustering result with the maximum total contour parameters has the optimal clustering effect, the corresponding K value is the optimal clustering K value, and the K scenes obtained in the steps from S2 to S6 are the clustering results according to the K value;
wherein, the step S8 specifically includes the following contents:
let s RDG sunrise force curves in the k scene, as shown in formula (6), and the output sequence P at the t moment in the scenew,t＝{Pwr,t,Pwo,t,...,Pws,t}; will sequence Pw,tThe output data is rearranged into a new sequence Y ═ Y from small to large1,y2,...,ysProcessing the sequence Y according to formula (7) to obtain YiBy the value z thereiniInstead of obtaining the sequence Z ═ Z1,z2,...,zs}; obtaining a new output interval [ minZ, maxZ ] based on the data in the sequence Z]；
Figure FDA0002837860680000032
Figure FDA0002837860680000033
Processing the 24-hour output data of the scene to obtain 24 groups of hour RDG output intervals
Figure FDA0002837860680000034
In the formula [ Prdg,i]Representing the output interval of the RDG in the scene at the ith moment of the scene; and sequentially carrying out interval construction on the K scenes to obtain 24 XK groups of hour output intervals.
2. The interval power flow typical scene construction method based on output curve aggregation as claimed in claim 1, wherein: the specific content of step S1 is:
providing wind speed and light intensity data v for a period of timei,t、hi,tWherein i is 1,2,3, …, L is the number of days in the period, 0 < L < 3650, t represents 24 hours; substituting the data into the formula (2) and the formula (3) for calculation;
Figure FDA0002837860680000041
in the formula Pwi,tIs as followsOutput power of wind power generation at the t-th moment of day i, vi,tRepresents the wind speed at the time t on the ith day, vs、vrAnd v0Respectively representing rated wind speed, cut-in wind speed and cut-out wind speed;
Figure FDA0002837860680000042
in the formula Pvi,tOutput power h of photovoltaic power generation at the ith momenti,tRepresents the actual illumination intensity at the t time of the ith day, hrIndicating the nominal light intensity, PVRRated output power;
calculating to obtain the time sequence data P of the wind power output of the ith dayW(i)＝(Pwi,1,Pwi,2,…,Pwi,24) And photovoltaic solar output time sequence data PV(i)＝(Pvi,1,Pvi,2,…,Pvi,24) The power curve psi (i) is combined as Pwi,1,Pwi,2,…,Pwi,24,Pvi,1,Pvi,2,…,Pvi,24]The sunrise force curves { ψ (1), ψ (2),.., ψ (L) } for L days are obtained.
3. The interval power flow typical scene construction method based on output curve aggregation as claimed in claim 1, wherein: the step S9 specifically includes the following steps:
step S91, calculating the power of the node:
Figure FDA0002837860680000051
in the formula SiRepresenting the node computation power, SLiInjecting power, y, into the nodei0Admittance for the terrestrial branch;
step S92, starting from the end of the distribution network, gradually advancing, and solving the power S of each branch of the distribution network by the node voltageijDistributing;
Figure FDA0002837860680000052
Figure FDA0002837860680000053
in the formula [ Sij]For branch power, Δ SijFor line power loss, Pj、QjFor node active and reactive power, ZijIs line impedance, SijIs the branch power;
step S93, calculating the node voltage interval from the branch power
Figure FDA0002837860680000054
Figure FDA0002837860680000055
Figure FDA0002837860680000056
In the formula
Figure FDA0002837860680000057
Respectively representing the longitudinal and transverse components of the branch pressure drop;
Figure FDA0002837860680000058
Figure FDA0002837860680000059
step S94, convergence criterion: the iteration termination criterion is the voltage interval [ U ] of each nodei]Relative upper and lower limitsThe deviation of the value in the last iteration is smaller than the allowable value as in equation (15);
Figure FDA0002837860680000061
in the formula
Figure FDA0002837860680000062
Respectively representing the upper limit and the lower limit of the voltage amplitude of the node j in the kth time, wherein epsilon is a deviation allowable value;
after convergence, the voltage interval [ U ] of each node is obtainedi]Namely, the interval power flow calculation result.
CN201910693437.0A 2019-07-30 2019-07-30 Interval power flow typical scene construction method based on output curve aggregation Active CN110336332B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910693437.0A CN110336332B (en) 2019-07-30 2019-07-30 Interval power flow typical scene construction method based on output curve aggregation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910693437.0A CN110336332B (en) 2019-07-30 2019-07-30 Interval power flow typical scene construction method based on output curve aggregation

Publications (2)

Publication Number Publication Date
CN110336332A CN110336332A (en) 2019-10-15
CN110336332B true CN110336332B (en) 2021-03-30

Family

ID=68148054

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910693437.0A Active CN110336332B (en) 2019-07-30 2019-07-30 Interval power flow typical scene construction method based on output curve aggregation

Country Status (1)

Country Link
CN (1) CN110336332B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110929399A (en) * 2019-11-21 2020-03-27 国网江苏省电力有限公司南通供电分公司 Wind power output typical scene generation method based on BIRCH clustering and Wasserstein distance
CN113914928A (en) * 2021-09-06 2022-01-11 中煤科工开采研究院有限公司 Support area dividing and accurate support method for fully mechanized coal mining face of coal mine

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107528350B (en) * 2017-09-28 2019-09-13 华中科技大学 A kind of wind power output typical scene generation method adapting to long -- term generation expansion planning
CN108242807B (en) * 2018-01-19 2019-10-22 广东电网有限责任公司河源供电局 A kind of reconstruction method of power distribution network containing photovoltaic considering multidimensional security constraint
CN108649605B (en) * 2018-05-22 2021-12-14 国网内蒙古东部电力有限公司通辽供电公司 DRE grid-connected access capacity planning method based on double-layer scene interval power flow

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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

Also Published As

Publication number Publication date
CN110336332A (en) 2019-10-15

Similar Documents

Publication Publication Date Title
CN108306303B (en) Voltage stability evaluation method considering load increase and new energy output randomness
CN110909911B (en) Aggregation method of multidimensional time series data considering space-time correlation
Saroha et al. Multi step ahead forecasting of wind power by genetic algorithm based neural networks
CN109086928A (en) Photovoltaic plant realtime power prediction technique based on SAGA-FCM-LSSVM model
CN110148068A (en) One kind being based on meteorological data similarity analysis and LSTM neural fusion photovoltaic plant ultra-short term power forecasting method
CN110110948B (en) Multi-target distributed power supply optimal configuration method
CN110336332B (en) Interval power flow typical scene construction method based on output curve aggregation
Alanazi et al. Long-term solar generation forecasting
JP2017135964A (en) Operation of photovoltaic power generation system
CN113128786B (en) Wind, light and water complementary capacity and site selection optimization method based on space grid search
CN112149905A (en) Photovoltaic power station short-term power prediction method based on wavelet transformation and wavelet neural network
CN109149656B (en) Gas-electricity interconnection comprehensive energy system unit combination method
CN113141008A (en) Data-driven power distribution network distributed new energy consumption capacity assessment method
CN104657796A (en) Distribution network planning method
CN108075471B (en) Multi-objective constraint optimization power grid scheduling strategy based on stochastic power output prediction
CN115021331A (en) Clean energy-based ground water wind-solar integrated compensation method based on contribution coefficient method
CN108694475B (en) Short-time-scale photovoltaic cell power generation capacity prediction method based on hybrid model
CN113052386B (en) Random forest algorithm-based distributed photovoltaic daily power generation amount prediction method and device
CN110309990B (en) New energy uncertainty planning method considering typical scene tolerance
Chicco et al. Characterization of solar irradiance profiles for photovoltaic system studies through data rescaling in time and amplitude
CN116128154A (en) Energy optimal configuration method and device for agricultural park comprehensive energy system
CN113312779B (en) High-satisfaction dynamic comprehensive planning method for low-carbon flexible power distribution network
Iliadis et al. Advanced energy management system based on PV and load forecasting for load smoothing and optimized peak shaving of islanded power systems
CN113962357A (en) GWO-WNN-based distributed photovoltaic power data virtual acquisition method
Özkan et al. A data mining-based wind power forecasting method: results for wind power plants in turkey

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