CN109378834A - Large scale electric network voltage stability margin assessment system based on information maximal correlation - Google Patents
Large scale electric network voltage stability margin assessment system based on information maximal correlation Download PDFInfo
- Publication number
- CN109378834A CN109378834A CN201811296116.9A CN201811296116A CN109378834A CN 109378834 A CN109378834 A CN 109378834A CN 201811296116 A CN201811296116 A CN 201811296116A CN 109378834 A CN109378834 A CN 109378834A
- Authority
- CN
- China
- Prior art keywords
- feature
- voltage stability
- formula
- correlation
- data
- 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.)
- Pending
Links
Classifications
-
- 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
- H02J3/24—Arrangements for preventing or reducing oscillations of power in networks
-
- 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]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
Large scale electric network voltage stability margin assessment system based on information maximal correlation, the system are primarily based on PMU data or software emulation, and obtaining one being capable of the approximate electric system large data sets for characterizing current all operation characteristics of electric system；On this basis, by repeatedly using MRMR algorithm, construction and voltage stability margin correlation maximum and the smallest feature set of redundancy；The function expression of relationship between selected variable and VSM is constructed using MATLAB；The real-time variable data obtained by electric system, can pass through expression formula direct estimation VSM.And result is fed back into spot dispatch personnel, to make a policy in time.The present invention can solve the disadvantages of Traditional measurements method low precision, speed are slow, poor to loss of data poor robustness and assessment analyticity, meet the real-time evaluation requirement of large-scale electrical power system.
Description
Technical field
The present invention relates to large-scale electrical power systems to run control technology field, specifically a kind of based on information maximal correlation
Large scale electric network voltage stability margin assessment system.
Background technique
Electric system is most complicated one of man-made system, and with China's expanding economy, the demand of network load is increasingly
Increase.And it is corresponding, the transmittability of electric system is not obviously improved, transmission capacity still be limited, system
Operating point is also increasingly closer to the limit of power system stability operation.Pass through the summary discovery to many large-scale blackouts, voltage
Collapse the entire power system collapse during from important role.Therefore, the voltage stability of electric system is carried out real
When monitoring, guarantee operation of power networks operating point have enough voltage stability margins, to the safe and reliable operation of entire electric system
It is significant.
With the extensive interconnection of current electric grid, the stable appraisal procedure of electric system of the tradition based on model needs to carry out big
The circuit approximately equivalent of scale, Evaluation accuracy is poor, while its estimating velocity is also difficult to meet the requirements.In recent years, with calculating
The intelligent algorithms such as the fast development of machine technology, artificial neural network and decision tree are introduced into the stable assessment of electric system.
Although there is these appraisal procedures faster estimating velocity also respectively to have deficiency in practical applications.Artificial neural network conduct
A kind of "black box" algorithm can not know the specific incidence relation between input and output, it is difficult to stable control and electricity for power grid
Failure backtracking after Force system unstability provides important information.Although decision tree class method analyticity is stronger, to loss of data
Robustness is very poor, and the loss of core feature data may cause the failure of entire assessment models.
Therefore, the stable appraisal procedure of voltage suitable for extensive interconnected electric power system how is constructed, system variable is provided
The Analytical Expression of " transparent " between the stable nargin of voltage improves the robustness problem to loss of data, final to realize to whole
Accurate, the quick and stable assessment of a Network Voltage Stability nargin is when previous urgent problem to be solved.
Summary of the invention
In order to solve problem above, the invention proposes a kind of large scale electric network voltage stabilizations based on information maximal correlation
Nargin assessment system, intend to solve Traditional measurements method low precision, speed it is slow, to loss of data poor robustness and assessment analyticity it is poor
The disadvantages of, meet the real-time evaluation requirement of large-scale electrical power system.
The technical scheme adopted by the invention is as follows:
Large scale electric network voltage stability margin assessment system based on information maximal correlation, comprising: basic data extract with
Memory module, static voltage stability characteristic extracting module, topological transformation module, online evaluation module；
The basic data is extracted and memory module, is used for based on scene/history PMU data, or using MATLAB and
Python program controls power system simulation software PSS/E, acquires the Operation of Electric Systems data under different operating point,
It is final to obtain the electric system large data sets for capableing of the approximate current all operation characteristics of electric system of characterization.Then for every
One operating point calculates its corresponding voltage stability margin VSM；
The static voltage stability characteristic extracting module is used to extract and voltage stability margin correlation maximum and superfluous
The feature set of remaining property minimum MRMR becomes for finding one group with target wherein containing the MRMR algorithm based on information redundancy
The characteristic variable collection of correlation maximum is measured, and the correlation that feature is mutual in variables set is minimum.MRMR algorithm is repeatedly used,
Multiple and different feature sets can then be obtained；
The topological transformation module is constructed between different variables and corresponding voltage stability margin by curve fitting technique
Mathematical analysis relationship, for each topology, record its acquisition voltage stability margin function expression and corresponding parameter, such as
Unknown change in topology is detected in fruit electric system, then according to newest current operating situation extract again characteristic variable collection and
Mathematical analysis relationship is constructed, new model to be assessed is generated.
The online evaluation module is used for the new variables data obtained using electric system, steady by the voltage of acquisition
Determine nargin expression formula and directly carry out the stable assessment of voltage, and result is fed back into spot dispatch personnel, to make a policy in time.
In the basic data extraction and memory module, using under MATLAB and Python programmed acquisition different operating point
Operation of Electric Systems data include: the distribution function that system randomly selects load parameter first, meanwhile, utilize optimal load flow OPF
Determine its dependent variable such as the distribution of reality/reactive power and position of tapping switch.Then, original generator/load distribution is generated.
Different generators/load increases direction, and identical original generator/load distribution is made to generate different voltage stability boundaries
Point.In order to record more electric system behaviors, increase knowledge base, the load growth rate of different zones is set as not by the present invention
Together, and in the same area keep power factor constant.Meanwhile increment load mainly pass through the same area generator carry out it is flat
Weighing apparatus.Increase direction the result is that increasing direction and being similar to actual load.Constantly to load/generated output power in system into
Row increases, until system reaches quiescent voltage collapse point POC.It is collected and stored in during this corresponding to all system operating points
System variable value.
The basic data is extracted with memory module, and voltage stability margin VSM calculation method is as follows:
Wherein: PPOCFor the maximum active power that current system can transmit, i.e. peak load at collapse point.POPTo work as
System active power under first operating status.VSM can characterize current point of operation to the distance between collapse of voltage point.
In the static voltage stability characteristic extracting module, MRMR algorithm, core concept is correlation between finding feature
The maximum value of difference between redundancy is realized especially by formula (2):
In formula: H (fi) and H (fj) be respectively ith feature and j-th of feature entropy, { f1, f2, f3…fFIt is wait mention
The feature set taken, F are characterized the number of features of concentration, MI (fi, y) and indicate fi" similar " degree between feature and target class y,
MI(fi, fj) indicate feature set internal feature between " similar " degree, specifically determined by following formula:
Wherein, fi,xFor fiX-th of element, yxFor x-th of element of target variable y, p (fI, x) and p (yx) respectively indicate
fI, xAnd yxMarginal probability density function, p (fi,x,yx) indicate their joint probability distribution, p (fi,x,fi,y) indicate fi,x
fi,yJoint probability distribution.
In the static voltage stability characteristic extracting module, the cumulative search routine of MRMR algorithm is as follows:
(a) first feature selecting.Association relationship between each candidate feature and target class is calculated by formula (3).
Data set (D) is created for optional features, and selects the highest characteristic (f of MI valuei, y) and it is used as first optional features.
(b) second feature selecting.Association relationship between first selected feature and other candidate features is counted by formula (4)
It obtains.Select MI (fi,fj) the smallest feature of value is as second selection feature.
(c) subsequent feature selecting.There are the data of the two selection features, correlation and redundancy point in data set D
It can not be calculated by formula (3) and formula (4).Hereafter the information redundancy optimal algorithm as shown in (2) formula is constantly applied, is selected item by item
Take subcharacter.If having reached the specified quantity of selected feature, algorithm is terminated.
The present invention provides a kind of large scale electric network voltage stability margin assessment system based on information maximal correlation, the system
It is primarily based on PMU data or software emulation, obtaining one being capable of the approximate electric power for characterizing current all operation characteristics of electric system
System large data sets；On this basis, by repeatedly using MRMR algorithm, construction and voltage stability margin correlation maximum and superfluous
The remaining the smallest feature set of property；The function expression of relationship between selected variable and VSM is constructed using MATLAB；Pass through electric system
The real-time variable data obtained, can pass through expression formula direct estimation VSM.And result is fed back into spot dispatch personnel, with
Just it makes a policy in time.
Compared with conventional method, beneficial effect of the present invention are as follows: system uses the appraisal procedure based on information redundance, energy
The comprehensive whole statistical probability considered between variable of enough systems, has better adaptability to extensive interconnected electric power system；System
System is excavated and is extracted to stable feature using maximal correlation minimal redundancy criterion, and over-fitting and poor fitting can be effectively prevented
The generation of phenomenon improves the precision of system evaluation；Based on the stability margin analytical expression that curve matching obtains, have good
Analyticity, can for the failure after the stable control and electric system unstability of power grid recall important information be provided；While method is most
The assessment models obtained eventually are convenient for storage and calculate, and have good estimating velocity.
Detailed description of the invention
Fig. 1 is system diagram of the invention；
Fig. 2 is the overview flow chart of the embodiment of the present invention；
Fig. 3 is the composition block diagram that basic data extracts with memory module in the present invention；
Fig. 4 is the composition block diagram of static voltage stability characteristic extracting module in the present invention；
Fig. 5 is the composition block diagram of topological transformation module in invention.
Specific embodiment
Large scale electric network voltage stability margin assessment system based on information maximal correlation, for large-scale power train
System carries out Transient Stability Evaluation, assesses in voltage stability margin of the global level to power grid entirety, guarantees system more steady
It is run in the state of fixed, reliable.The assessment system is extracted it is characterized by comprising basic data and memory module, quiescent voltage are steady
Determine characteristic extracting module, four parts of topological transformation module and online evaluation module.
Basic data is extracted and memory module, is based on scene/history PMU data, or utilize MATLAB and Python program
Power system simulation software PSS/E is controlled, the Operation of Electric Systems data under different operating point is acquired, finally obtains one
A electric system large data sets for capableing of the approximate current all operation characteristics of electric system of characterization.Then for each work
Point calculates its corresponding voltage stability margin VSM.
Using the Operation of Electric Systems data under MATLAB and Python programmed acquisition different operating point, detailed process is such as
Under: system randomly selects the distribution function of load parameter first, meanwhile, reality/reactive power is determined using optimal load flow (OPF)
Its dependent variable such as distribution and position of tapping switch.Then, original generator/load distribution is generated.Different generator/loads
Increase direction, identical original generator/load distribution is made to generate different voltage stability boundary points.In order to record more electricity
Force system behavior increases knowledge base, and the load growth rate of different zones is set as different by the present invention, and is kept in the same area
Power factor is constant.Meanwhile increment load mainly passes through the generator of the same area and is balanced.The result is that it is similar to increase direction
Increase direction in actual load.Constantly load/generated output power in system is increased, until system arrival is quiet
State collapse of voltage point (POC).It is collected and stored in system variable value corresponding to all system operating points during this.
The calculating works of voltage stability margin are as follows:
Wherein PPOCFor the maximum active power that current system can transmit, i.e. peak load at collapse point.POPFor in the ban
System active power under operating status.VSM can characterize current point of operation to " distance " between collapse of voltage point.
Static voltage stability characteristic extracting module: it extracts and voltage stability margin correlation maximum and redundancy minimum
(MRMR) feature set.Wherein contain the MRMR algorithm based on information redundancy, it is intended to find one group it is related to target variable
Property maximum characteristic variable collection, and the correlation that feature is mutual in variables set is minimum.Repeatedly use MRMR algorithm, then it can be with
Obtain multiple and different feature sets.
MRMR algorithm, core concept are the maximum values of difference between correlation and redundancy between finding feature, specific logical
Cross formula (2) realization:
H (f in formulai) and H (fj) be respectively ith feature and j-th of feature entropy, { f1, f2, f3…fFIt is to be extracted
Feature set, F is characterized the number of features of concentration, MI (fi, y) and indicate fi" similar " degree between feature and target class y, MI
(fi, fj) indicate feature set internal feature between " similar " degree, specifically determined by following formula:
Wherein, fi,xFor fiX-th of element, yxFor x-th of element of target variable y, p (fI, x) and p (yx) respectively indicate
fI, xAnd yxMarginal probability density function, p (fi,x,yx) indicate their joint probability distribution, p (fi,x,fi,y) indicate fi,x
fi,yJoint probability distribution.
MRMR algorithm is a kind of incremental search process, and referred to as " order algorithm ", specifically cumulative search routine is as follows:
(a) first feature selecting.Association relationship between each candidate feature and target class is calculated by formula (3).
Data set (D) is created for optional features, and selects the highest characteristic (f of MI valuei, y) and it is used as first optional features.
(b) second feature selecting.Association relationship between first selected feature and other candidate features is counted by formula (4)
It obtains.Select MI (fi,fj) the smallest feature of value is as second selection feature.
(c) subsequent feature selecting.There are the data of the two selection features, correlation and redundancy point in data set D
It can not be calculated by formula (3) and formula (4).Hereafter the information redundancy optimal algorithm as shown in (2) formula is constantly applied, is selected item by item
Take subcharacter.If having reached the specified quantity of selected feature, algorithm is terminated.
Topological transformation module: the mathematics between different variables and corresponding voltage stability margin is constructed by curve fitting technique
Parsing relationship.
Topological transformation module: for each topology, the voltage stability margin function expression of its acquisition and corresponding is recorded
Parameter.If detecting unknown change in topology in electric system, feature is extracted according to newest current operating situation again
Variables set and building mathematical analysis relationship, generate new model to be assessed.
Online evaluation module: the new variables data obtained using electric system can pass through the voltage stability margin of acquisition
Expression formula directly carries out the stable assessment of voltage, and result is fed back to spot dispatch personnel, to make a policy in time.
Embodiment:
General frame in embodiment is as shown in Figure 1, system is mainly extracted by basic data and memory module, quiescent voltage are steady
Determine characteristic extracting module, topological transformation module and online evaluation module composition.Basic data is extracted and memory module, the purpose is to
Large-scale history data set based on live PMU establishes the statistical model for being similar to all stochastic variable probability distribution；It is static
Voltage stabilization characteristic extracting module constructs and voltage stability margin correlation maximum and the smallest feature set of redundancy；Topology
In conversion module, by curve matching, the topological relation between different variables and corresponding VSM is constructed；Online evaluation module, passes through
The new variables data that electric system obtains, can feed back to spot dispatch people by expression formula direct estimation VSM, and by result
Member, to make a policy in time.
The overall execution flow chart of embodiment is as shown in Fig. 2, specifically include following steps:
(1) PSS/E emulator is controlled using MATLAB and Python program, automatic collection PSS/E data, is established
The electric system large data sets of one record performance variable and corresponding VSM, and calculated corresponding to each operating point by formula (1)
Voltage stability margin VSM.
(2) by MRMR algorithm, the relationship between performance variable and VSM is constructed, and record the table of heterogeneous networks topology
Up to formula.
(3) when new service condition occurs, compare present topology and data concentrate existing topology, determine that the topology is
It is no to be recorded.
(3 ') use corresponding feature set and expression formula if the topology is existing topological classification in data set.
(3 ") if the topology be new topology MRMR process is applied in new topology under new service condition,
Obtain new expression formula.
(4) according to obtained relationship, voltage stability margin assessment is carried out to current operating point.
Fig. 3 is the composition block diagram that basic data extracts with memory module in the present invention.In the module, system is random first
The distribution function of load parameter is chosen, meanwhile, the distribution of reality/reactive power and tap switch are determined using optimal load flow (OPF)
Its dependent variable such as position.Then, original generator/load distribution is generated.Different generators/load increases direction, makes phase
Same original generator/load distribution generates different voltage stability boundary points.In order to record more electric system behaviors, increase
Add knowledge base, the load growth rate of different zones is set as different by the present invention, and keeps power factor constant.Meanwhile increment is negative
The generator that lotus mainly passes through the same area is balanced.Increase direction the result is that increasing direction and being similar to actual load.No
It is disconnected that load/generated output power in system is increased, until system reaches quiescent voltage collapse point.It collects and stores
System variable value corresponding to all system operating points in the process.
Fig. 4 is the composition block diagram of static voltage stability characteristic extracting module in the present invention.Specific workflow are as follows:
(1) correlation calculations.Between correlation and feature and target class y by calculating data set D and target class y
" similar " degree, finds out the feature with target variable correlation maximum.
(2) redundant computation.By calculating selected characteristic value redundancy that may be present, filter out with target variable redundancy most
Small feature.
(3) MRMR algorithm.By MRMR algorithm, the feature of one group with target variable correlation maximum is found, and they it
Between correlation it is minimum.
As a preferred solution of the present invention, relative parameters setting is as follows in the present embodiment:
(1) in PSS/E system, test and generate the operating point 8168 of a variety of different voltage margins, each work
Contain 31838 variables in point.
(2) in 8168 samples of generation, wherein 6535 samples are trained for random selection, and 1633 samples give over to
Test.
(3) then 10 MRMR feature sets (M=10) of parallel generation, include that 5 features become in each MRMR feature set
It measures (N=5).
Fig. 5 is the composition block diagram of topological transformation module in the present invention.Specific workflow are as follows:
(1) by MRMR algorithm, the relationship between performance variable and VSM is constructed, and record the table of heterogeneous networks topology
Up to formula.
(2) when new service condition occurs, compare present topology and data concentrate existing topology, determine that the topology is
It is no to be recorded.
(3) if the topology is existing topology in data set, corresponding feature set and expression formula are used.
(4) if the topology is that MRMR algorithm is applied in new topology by new topology under new service condition,
Obtain new expression formula.
As a preferred solution of the present invention, the characteristic variable type in first MRMR feature set finally obtained,
And the mathematical function relationship between each characteristic variable and voltage stability margin is as shown in the table；
As a preferred solution of the present invention, in online evaluation module, due to that can have been generated during multiple MRMR
The feature set of 10 different solutions, each selection can provide VSM estimated value, finally be capable of providing more acurrate result
Aggregation decisions.
The statistical accuracy that the present invention is assessed in 1648 bus systems test result is as follows table:
The relevant measuring and calculation time is as shown in the table:
Dependence test the result shows that, if encountering new topological structure in 1648 bus-bar systems, the online updating stage can
To provide accurate assessment result in 10 minutes.If the topology of electric system is recorded in known topology List, program
Assessment result can be provided in 0.67 second.It is fully able to meet the requirement of online evaluation.
The final testing result of embodiment shows more effectively, more completely indicate target category information with MRMR method.
Meanwhile compared with other transparent AIS tools, MRMR method has higher assessment accuracy rate.It is unseen before encountering to open up
When flutterring variation, MRMR technology can be used the example newly obtained and refresh selected variable and corresponding function expression.The side MRMR
Method has quick training and predetermined speed, can satisfy the requirement that real-time online uses.In practical applications, it is recommended to use sample
It is equal to or more than the variable quantity of the determining parameter of experiment in this amount, the quantity and each feature set of parallel MRMR process.
Claims (10)
1. the large scale electric network voltage stability margin assessment system based on information maximal correlation, characterized by comprising: basic number
According to extraction and memory module, static voltage stability characteristic extracting module, topological transformation module, online evaluation module；
The basic data is extracted and memory module, is used for based on scene/history PMU data, or using MATLAB and
Python program controls power system simulation software PSS/E, acquires the Operation of Electric Systems data under different operating point,
It is final to obtain the electric system large data sets for capableing of the approximate current all operation characteristics of electric system of characterization；Then for every
One operating point calculates its corresponding voltage stability margin VSM；
The static voltage stability characteristic extracting module is used to extract and voltage stability margin correlation maximum and redundancy
The feature set of minimum MRMR, wherein the MRMR algorithm based on information redundancy is contained, for finding one group and target variable phase
The maximum characteristic variable collection of closing property, and the correlation that feature is mutual in variables set is minimum；MRMR algorithm is repeatedly used, then can
Access multiple and different feature sets；
The topological transformation module constructs the mathematical analysis relationship between different variables and corresponding voltage stability margin, for every
A topology records voltage stability margin function expression and corresponding parameter that it is obtained, if detected not in electric system
The change in topology known then extracts characteristic variable collection and building mathematical analysis relationship according to newest current operating situation again, raw
The model to be assessed of Cheng Xin；
The online evaluation module is used for the new variables data obtained using electric system, abundant by the voltage stabilization of acquisition
Degree expression formula directly carries out the stable assessment of voltage, and result is fed back to spot dispatch personnel.
2. the large scale electric network voltage stability margin assessment system based on information maximal correlation according to claim 1, special
Sign is: in the basic data extraction and memory module, using under MATLAB and Python programmed acquisition different operating point
Operation of Electric Systems data include: the distribution function that system randomly selects load parameter first, meanwhile, utilize optimal load flow OPF
Determine its dependent variable such as the distribution of reality/reactive power and position of tapping switch；Then, original generator/load distribution is generated；
Different generators/load increases direction, and identical original generator/load distribution is made to generate different voltage stability boundaries
Point；The load growth rate of different zones is set as different by the system, and keeps power factor constant in the same area；Meanwhile
The generator that increment load mainly passes through the same area is balanced；The result is that increasing direction is similar to actual load increase side
To；Constantly load/generated output power in system is increased, until system reaches quiescent voltage collapse point POC；It receives
Collect and stores system variable value corresponding to all system operating points in the process.
3. the large scale electric network voltage stability margin assessment system based on information maximal correlation according to claim 1, special
Sign is: the basic data is extracted with memory module, and voltage stability margin VSM calculation method is as follows:
Wherein: PPOCFor the maximum active power that current system can transmit, i.e. peak load at collapse point；POPTo transport in the ban
System active power under row state；VSM can characterize current point of operation to the distance between collapse of voltage point.
4. the large scale electric network voltage stability margin assessment system based on information maximal correlation according to claim 1, special
Sign is: in the static voltage stability characteristic extracting module, MRMR algorithm, core concept be between finding feature correlation with
The maximum value of difference between redundancy is realized especially by formula (2):
In formula: H (fi) and H (fj) be respectively ith feature and j-th of feature entropy, { f1, f2, f3…fFIt is to be extracted
Feature set, F are characterized the number of features of concentration, MI (fi, y) and indicate fi" similar " degree between feature and target class y, MI
(fi, fj) indicate feature set internal feature between " similar " degree, specifically determined by following formula:
Wherein, fi,xFor fiX-th of element, yxFor x-th of element of target variable y, p (fI, x) and p (yx) respectively indicate fI, x
And yxMarginal probability density function, p (fi,x,yx) indicate their joint probability distribution, p (fi,x,fi,y) indicate fi,x fi,y's
Joint probability distribution.
5. the large scale electric network voltage stability margin assessment system based on information maximal correlation according to claim 1, special
Sign is: in the static voltage stability characteristic extracting module, the cumulative search routine of MRMR algorithm is as follows:
(a) first feature selecting；Association relationship between each candidate feature and target class is calculated by formula (3)；For choosing
Fixed characteristic creates data set (D), and selects the highest characteristic (f of MI valuei, y) and it is used as first optional features；
(b) second feature selecting；Association relationship between first selected feature and other candidate features is calculated by formula (4)
It arrives；Select MI (fi,fj) the smallest feature of value is as second selection feature；
(c) subsequent feature selecting；There are the data of the two selection features, the correlation and redundancy in data set D respectively can
It is calculated by formula (3) and formula (4)；Hereafter the information redundancy optimal algorithm as shown in (2) formula is constantly applied, chooses son item by item
Feature；If having reached the specified quantity of selected feature, algorithm is terminated.
6. the large scale electric network voltage stability margin appraisal procedure based on information maximal correlation, it is characterised in that including following step
It is rapid:
(1), it is based on scene/history PMU data, or using MATLAB and Python program to power system simulation software PSS/E
It is controlled, acquires the Operation of Electric Systems data under different operating point, final acquisition one approximate can characterize current electric power
The electric system large data sets of all operation characteristics of system；Then it is directed to each operating point, calculates its corresponding voltage stabilization
Nargin VSM；
(2), by repeatedly using MRMR algorithm, in electric system large data sets, M different feature sets, Mei Gete are extracted
Sign be concentrated with N number of variable, the voltage stability margin correlation maximum of these variables and system, and the feature in same feature set it
Between correlation it is minimum；Specific MRMR specific algorithm includes correlation calculations, and redundancy calculates and MRMR calculates three parts；
(3), the function expression of relationship between selected variable and VSM is constructed using the curve fitting module in MATLAB；
(4), when detecting that system topological changes, then will occur new parsing relationship, i.e., existing table between variable and VSM
It is no longer applicable in up to formula；It needs to apply to MRMR process in new topology under new service condition at this time, obtains variable and VSM
New relationship, realize online data update；
(5), the new variables data obtained by electric system, can pass through expression formula direct estimation VSM；And result is fed back to
Spot dispatch personnel, to make a policy in time.
7. the large scale electric network voltage stability margin appraisal procedure based on information maximal correlation according to claim 6, special
Sign is: most relevance minimal redundancy (MRMR) algorithm in the step (2), target is looked on the basis of mutual information
To one group of relevant complementary characteristic；The basis of this technology is, if having close connection between two features, they are dividing
Similar effect is played in class or prediction；Both therefore, there is no need to will be included in optional features concentrates, and whether they are from no
It is highly relevant with target object class；MRMR algorithm includes that relevance calculating and redundancy calculate two parts；Specifically, electric system
Data set D is expressed as S sample and F={ f1, f2, f3…fFFeature (operating parameter)；Y is target variable VSM；MRMR algorithm
Target is to find subspace M={ m1, m2, m3….mM, RM；From F and RFTarget y class can be complementally portrayed comprehensively.
8. the large scale electric network voltage stability margin appraisal procedure based on information maximal correlation according to claim 6, special
Sign is: the method for the correlation calculations in the step (2) is as follows:
The correlation (being denoted as U (D, y)) of data set D and target class y is by all mutual informations between each feature and target class y
(MI) average value of value is measured, as follows:
Wherein F is characterized the size of collection；MI(fi, y) and indicate " similar " degree between feature and target class y, it is defined as follows；
Wherein, fi,xFor fiX-th of element, yxFor x-th of element of target variable y, p (fI, x) and p (yx) respectively indicate fI, x
And yxMarginal probability density function, p (fi,x,yx) indicate their joint probability distribution；
The method that redundancy in the step (2) calculates is as follows:
It there may be redundancy between the feature chosen according to formula (2)；V (F) indicates similar (redundancy) journey between different characteristic
Degree, can be calculated by formula (4):
Wherein fi, fjRespectively indicate ith and jth feature, MI (fi,fj) calculated by formula (5), indicate their mutual information；
Wherein, p (fi,x,fi,y) indicate fi,x fi,yJoint probability distribution；
The method that final MRMR in the step (2) is calculated is as follows:
The target of maximal correlation minimal redundancy algorithm is the feature (VSM) for finding one group with target variable correlation maximum, and root
According to expression above (2) and (4), the correlation between them is minimum；In practice, most relevance U (D, y) and minimal redundancy
V (F) cannot be realized always simultaneously；It is optimized, is combined into a standard, provided by expression formula (6)；Wherein, H
(fi) and H (fj) be respectively ith feature and j-th of feature entropy；
9. the large scale electric network voltage stability margin appraisal procedure based on information maximal correlation according to claim 6, special
Sign is: the MRMR algorithm in the step (2) is a kind of incremental search process, referred to as " order algorithm ", is added up and following institute
Show:
A) first feature selecting；Association relationship between each candidate feature and target class is calculated by formula (3)；It is selected
Characteristic create data set (D), and select the highest characteristic (f of MI valuei, y) and it is used as first optional features；
B) second feature selecting；Association relationship between first selected feature and other candidate features is calculated by formula (5)
It arrives；Select MI (fi,fj) the smallest feature of value is as second selection feature；
C) subsequent feature selecting；There are the data of the two selection features, the correlation and redundancy in data set D are respectively by formula
(2) it is calculated with formula (4)；Maximal correlation minimal redundancy criterion is realized, expression formula is formula (6)；It should by reusing
Criterion chooses subcharacter item by item；If having reached the specified quantity of selected feature, algorithm is terminated.
10. the large scale electric network voltage stability margin appraisal procedure based on information maximal correlation according to claim 6, special
Sign is: in the step (5), final evaluation result R is calculated using formula (7)F；
M is the sum of selected characteristic set, and N is the variable number in each characteristic set, RijIt is j-th in ith feature set
The corresponding estimated result of characteristic variable.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811296116.9A CN109378834A (en) | 2018-11-01 | 2018-11-01 | Large scale electric network voltage stability margin assessment system based on information maximal correlation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811296116.9A CN109378834A (en) | 2018-11-01 | 2018-11-01 | Large scale electric network voltage stability margin assessment system based on information maximal correlation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109378834A true CN109378834A (en) | 2019-02-22 |
Family
ID=65391071
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811296116.9A Pending CN109378834A (en) | 2018-11-01 | 2018-11-01 | Large scale electric network voltage stability margin assessment system based on information maximal correlation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109378834A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814394A (en) * | 2020-06-30 | 2020-10-23 | 三峡大学 | Power system safety assessment method based on correlation and redundancy detection |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140207764A1 (en) * | 2013-01-21 | 2014-07-24 | International Business Machines Corporation | Dynamic feature selection with max-relevancy and minimum redundancy criteria |
| US20140207799A1 (en) * | 2013-01-21 | 2014-07-24 | International Business Machines Corporation | Hill-climbing feature selection with max-relevancy and minimum redundancy criteria |
| CN107171315A (en) * | 2017-05-18 | 2017-09-15 | 广东电网有限责任公司电力科学研究院 | A kind of transient stability evaluation in power system method based on RPTSVM |
| CN108667007A (en) * | 2018-04-16 | 2018-10-16 | 清华大学 | Meter and the voltage stability margin computational methods of electric-gas coupled system constraint |
-
2018
- 2018-11-01 CN CN201811296116.9A patent/CN109378834A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140207764A1 (en) * | 2013-01-21 | 2014-07-24 | International Business Machines Corporation | Dynamic feature selection with max-relevancy and minimum redundancy criteria |
| US20140207799A1 (en) * | 2013-01-21 | 2014-07-24 | International Business Machines Corporation | Hill-climbing feature selection with max-relevancy and minimum redundancy criteria |
| CN107171315A (en) * | 2017-05-18 | 2017-09-15 | 广东电网有限责任公司电力科学研究院 | A kind of transient stability evaluation in power system method based on RPTSVM |
| CN108667007A (en) * | 2018-04-16 | 2018-10-16 | 清华大学 | Meter and the voltage stability margin computational methods of electric-gas coupled system constraint |
Non-Patent Citations (1)
| Title |
|---|
| YOUPING FAN; XIN LI; PENG ZHANG: "Real-Time Static Voltage Stability Assessment in Large-Scale Power Systems Based on Maximum-Relevance Minimum-Redundancy Ensemble Approach", 《IEEE》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814394A (en) * | 2020-06-30 | 2020-10-23 | 三峡大学 | Power system safety assessment method based on correlation and redundancy detection |
| CN111814394B (en) * | 2020-06-30 | 2023-08-25 | 三峡大学 | Power system safety assessment method based on correlation and redundancy detection |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mosadeghy et al. | A time-dependent approach to evaluate capacity value of wind and solar PV generation | |
| CN103488869A (en) | Wind power generation short-term load forecast method of least squares support vector machine | |
| CN111628494B (en) | Low-voltage distribution network topology identification method and system based on logistic regression method | |
| TWI540533B (en) | Method of short-term wind power generation forecasting | |
| Cui et al. | Ambient synchrophasor measurement based system inertia estimation | |
| CN112182720B (en) | Building energy consumption model evaluation method based on building energy management application scene | |
| CN109378835A (en) | Based on the large-scale electrical power system Transient Stability Evaluation system that mutual information redundancy is optimal | |
| CN112994097A (en) | High-proportion distributed photovoltaic cooperative control method based on intelligent distribution transformer terminal system | |
| CN111242161A (en) | Non-invasive non-resident user load identification method based on intelligent learning | |
| CN104578053A (en) | Power system transient stability prediction method based on disturbance voltage trajectory cluster features | |
| CN115859099A (en) | Sample generation method and device, electronic equipment and storage medium | |
| CN103617447A (en) | Evaluation system and method for intelligent substation | |
| CN111582630A (en) | Method and system for determining low-voltage transformer area line loss rate evaluation value | |
| CN112446175B (en) | Distributed power supply site selection optimization method based on power distribution network probability load flow calculation | |
| CN109378834A (en) | Large scale electric network voltage stability margin assessment system based on information maximal correlation | |
| CN105701265A (en) | Double-fed wind generator modeling method and apparatus | |
| CN109684749B (en) | Photovoltaic power station equivalent modeling method considering operating characteristics | |
| CN111965442A (en) | Energy internet fault diagnosis method and device under digital twin environment | |
| CN116502894A (en) | Photovoltaic transformer area power failure risk assessment method and device, electronic equipment and storage medium | |
| CN115296295B (en) | Low-voltage distribution transformer area topology identification method and system based on cluster analysis and graph theory | |
| CN116911161A (en) | Data-enhanced deep learning transient voltage stability evaluation method | |
| CN114240003A (en) | New energy output prediction method, system, storage medium and equipment | |
| CN115598459A (en) | Power failure prediction method for 10kV feeder line fault of power distribution network | |
| Menke | A comprehensive approach to implement monitoring and state estimation in distribution grids with a low number of measurements | |
| CN115809761B (en) | Voltage quality analysis method and system based on low-voltage transformer area |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190222 |
|
| RJ01 | Rejection of invention patent application after publication |