CN108988347A - A kind of adjusting method and system that power grid Transient Voltage Stability sample set classification is unbalance - Google Patents
A kind of adjusting method and system that power grid Transient Voltage Stability sample set classification is unbalance Download PDFInfo
- Publication number
- CN108988347A CN108988347A CN201810866566.0A CN201810866566A CN108988347A CN 108988347 A CN108988347 A CN 108988347A CN 201810866566 A CN201810866566 A CN 201810866566A CN 108988347 A CN108988347 A CN 108988347A
- Authority
- CN
- China
- Prior art keywords
- sample
- forecast
- unstability
- power grid
- collection
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 230000001052 transient effect Effects 0.000 title claims abstract description 38
- 238000003066 decision tree Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000004088 simulation Methods 0.000 claims abstract description 29
- 238000013097 stability assessment Methods 0.000 claims abstract description 10
- 238000012544 monitoring process Methods 0.000 claims abstract description 8
- 238000012797 qualification Methods 0.000 claims abstract description 8
- 238000012360 testing method Methods 0.000 claims description 24
- 238000007689 inspection Methods 0.000 claims description 17
- 238000003786 synthesis reaction Methods 0.000 claims description 13
- 230000015572 biosynthetic process Effects 0.000 claims description 8
- 230000001105 regulatory effect Effects 0.000 claims description 4
- 230000000717 retained effect Effects 0.000 claims description 3
- 230000005611 electricity Effects 0.000 claims 3
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000011156 evaluation Methods 0.000 abstract description 2
- 238000007418 data mining Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000002790 cross-validation Methods 0.000 description 2
- 230000007812 deficiency Effects 0.000 description 2
- 230000006641 stabilisation Effects 0.000 description 2
- 238000011105 stabilization Methods 0.000 description 2
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
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
- H02J3/12—Circuit arrangements for ac mains or ac distribution networks for adjusting voltage in ac networks by changing a characteristic of the network load
-
- 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Power Engineering (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses adjusting methods and system that power grid Transient Voltage Stability sample set classification is unbalance, belong to Power System Stability Analysis evaluation areas.Method of the invention is using the historical sample collection collected from power grid historical record as the chief component of training sample set, using under management and running plan forecast sample generated to the time-domain-simulation of forecast failure collection, and is relatively tested to the qualification of forecast sample using the Euclidean distance between different classes of sample；Forecast sample is merged with historical sample, constitutes the training sample set for excavating study, realizes alleviation and adjusting to historical sample collection classification unbalance；Classification learning is carried out to training sample set using decision Tree algorithms, decision-tree model is obtained, using decision-tree model as the power grid Transient Voltage Stability assessment models during real-time monitoring.
Description
Technical field
The invention belongs to Power System Stability Analysis evaluation areas, in particular to a kind of power grid Transient Voltage Stability sample set
The adjusting method and system that classification is unbalance.
Background technique
With the fast development of the technologies such as data mining and machine learning, has in power grid much use data mining at present
Method carries out on-line monitoring and the assessment of power grid Transient Voltage Stability.Traditional monitoring of power grid Transient Voltage Stability and assessment master
To obtain the training sample of data mining by the offline time-domain-simulation of power grid, but consolidating due to power network modeling and time-domain-simulation
There is error, the problem of reliability deficiency may be brought by the training sample that time-domain-simulation generates.To improve training sample data
The reliability in source, can be from historical sample of the collection power grid under various historical failures in power grid history log.However, power grid
It can maintain Transient Voltage Stability in most cases in actual operation, the case of unstability is relatively fewer, this will be caused from history
The classification serious unbalance for the historical sample collection collected in record.If not doing any processing to classification unbalance, excavation will lead to
It is too low to the attention rate of unstability sample in training process, to cause finally obtained power grid Transient Voltage Stability assessment models pair
The recalling of unstability sample is spent low.
Has more mature management and running plan arrangement method in dispatching of power netwoks operation platform, if management and running meter can be utilized
It draws, time-domain-simulation is carried out to the various forecast failures of the following power grid in a short time, it is obtained pre- that time-domain-simulation will be greatly improved
The reliability of test sample sheet.It, can be to above-mentioned historical sample collection if suitably synthesizing the forecast sample of certain amount by this method
Classification unbalance effectively alleviated and adjusted.
Patent document CN105139289A discloses a kind of " power grid Transient Voltage Stability based on wrong point cost classification learning
Appraisal procedure ", this method are constituted based on the dynamic measurement data of synchronous phasor measurement unit from a large amount of dynamic measurement data
Time series in extract the crucial subsequence closely related with electric network state；Pass through the setting stabilization of power grids, instability status
Different mistakes divide cost, introduce weight coefficient to learning sample；Classified using the decision Tree algorithms for incorporating sample weights coefficient
Study, obtains decision-tree model, decision-tree model is used to monitor on-line, implements to assess to power grid Transient Voltage Stability situation,
The technical solution, which is laid particular emphasis on, directly carries out classification learning to the data set of class imbalance using cost sensitive learning method.
Summary of the invention
In view of the deficiencies of the prior art, the purpose of the present invention is intended to provide a kind of power grid Transient Voltage Stability sample set classification
Unbalance adjusting method recalls ability to enhance power grid Transient Voltage Stability assessment models to unstability sample.
To achieve the above object, the present invention adopts the following technical scheme:
The beneficial effects of the present invention are:
A kind of adjusting method that power grid Transient Voltage Stability sample set classification is unbalance, including
S1, historical sample collection S is collected from power grid history logp0；
S2, by generating unstability forecast sample, all unstabilitys to the time-domain-simulation of forecast failure under management and running plan
Forecast sample forms an initial predicted sample set Sp0；
S3, the historical sample collection S in step S1 is utilizedp0To the initial predicted sample set S in step S2p0In lost
Steady forecast sample carries out approval test, and underproof unstability forecast sample is rejected, all through examining qualified unstability to predict
Sample forms qualified forecast sample collection Sp1；
S4, the qualified forecast sample collection S obtained using step S3p1To adjust historical sample collection S in above-mentioned steps S10, will
Sp1It is merged into S0In, a training sample set is formed, to overcome classification unbalance；；
S5, classification learning is carried out using training sample set of the decision Tree algorithms to step S4, obtains a decision-tree model,
Using decision-tree model as the Transient Voltage Stability assessment models of power grid, the Transient Voltage Stability state of power grid is supervised in real time
It surveys and assesses.
The step S1 specifically:
Historical failure collection, node collection and the characteristic variable collection of power grid are obtained from the history log of power grid, and are collected
Under each historical failure in power grid the characteristic variable value of each node and power grid operating status Z, power grid is in steady operational status
It is denoted as Z=1, power grid is in unstability operating status and is denoted as Z=-1, and the data acquisition system collected under a historical failure is gone through at one
History sample, total N0A historical sample is integrated into a historical sample collection S0, it is total to stablize historical sample in statistical history sample set respectively
Number N1With unstability historical sample sum N-1, wherein N1+N-1=N0。
The step S2 specifically:
From the management and running platform of power grid obtain present period power grid node collection, characteristic variable collection, it is n hours following in
Management and running plan and n hours future in forecast failure collection, using computer time-domain simulation method to the power grid not
Carry out the various forecast failures in n hours under management and running plan and carries out NpSecondary time-domain-simulation records each time-domain-simulation mistake respectively
In journey in power grid each characteristic variable value and power grid of each node after meeting with failure operating status Z, wherein Z=1 represents power grid
In steady operational status, Z=-1 represents power grid and is in unstability operating status, judges Z, if Z=-1, when by this
One unstability forecast sample of the Data Synthesis recorded in the simulation process of domain does not make the synthesis of unstability forecast sample if Z=1, system
Count the unstability forecast sample number N of all synthesisp0, Np0A unstability forecast sample forms an initial predicted sample set Sp0。
The step S3 includes:
S31, the historical sample collection S in step S1 is utilized0With the initial predicted sample set S in step S2p0Synthesize an inspection
Test sample set S1, wherein test samples sum is Nt=N0+Np0；
The initial predicted sample set S of S32, any selecting step S2p0In unstability forecast sample i, and going through from step S1
History sample set S0All N1It is arbitrarily chosen in a stable historical sample and stablizes historical sample j, calculate the Europe between the two samples
Family name distance d (i, j), wherein 1≤i≤Np0, 1≤j≤N1；
S33, the test samples collection S for successively calculating step S311In all NtA test samples are predicted with unstability in step S32
Euclidean distance between sample i unstability forecast sample i, all N that will be calculatedtThe maximum value of a Euclidean distance numerical value is denoted as
d(i,Nt)；
S34, test samples collection S in step S31 is successively calculated1All NtStablize history in a test samples and step S32
Euclidean distance between sample j, all N that will be calculatedtThe maximum value of a Euclidean distance numerical value is denoted as d (j, Nt)；
S35, the obtained Euclidean distance of step S32~S34 is compared and is judged, if d (i, Nt) >=d (i, j) and d
(j,Nt) >=d (i, j) then illustrates that unstability forecast sample i is unqualified, step S37 is carried out, if d (i, Nt) < d (i, j) and d (j,
Nt) >=d (i, j) or d (i, Nt) >=d (i, j) and d (j, Nt) < d (i, j) or d (i, Nt) < d (i, j) and d (j, Nt) <
D (i, j) then illustrates that unstability forecast sample i is qualified in this inspection, carries out step S36；
S36, traversal history sample set S0In all N1A stable historical sample, repeat the above steps S32~S35, obtains
The inspection result of unstability forecast sample i；
S37, traversal initial predicted sample set Sp0In all Np0A unstability forecast sample, the S32 that repeats the above steps~
S36 obtains all Np0The inspection result of a unstability forecast sample；
All N that S38, read step S37 are obtainedp0The inspection result of a unstability forecast sample, if unstability forecast sample is examined
Test result be it is unqualified, then by the unstability forecast sample from initial predicted sample set Sp0Middle rejecting, if unstability forecast sample is examined
It as a result is qualification, then by the unstability forecast sample in initial predicted sample set Sp0In retained, count qualified unstability prediction
Total sample number Np1, all Np1A qualified unstability forecast sample remained constitutes qualified forecast sample collection Sp1。
Correspondingly, the present invention also provides a kind of regulating system that power grid Transient Voltage Stability sample set classification is unbalance, packets
It includes:
Historical sample collection generation module is used to collect historical sample from power grid history log, be gone through with spanning set
History sample set Sp0；
Initial predicted sample set generation module is used for by raw to the time-domain-simulation of forecast failure under management and running plan
At unstability forecast sample, all unstability forecast samples form an initial predicted sample set Sp0；
Qualified forecast sample collection generation module passes through historical sample collection generation module historical sample collection S generatedp0It is right
Initial predicted sample set generation module initial predicted sample set S generatedp0In all unstability forecast samples carry out qualification
It examines, underproof unstability forecast sample is rejected, it is all through examining qualified unstability forecast sample to form qualified forecast sample
Collect Sp1；
Training sample set generation module passes through the qualified forecast sample obtained using qualified forecast sample collection generation module
Collect Sp1To adjust the historical sample collection S of above-mentioned steps initial predicted sample set generation module generation0, by Sp1It is merged into S0In, shape
At a training sample set；
Decision-tree model generation module, the training sample for using decision Tree algorithms to generate training sample set generation module
Collection carries out classification learning, obtains a decision-tree model, using decision-tree model as the Transient Voltage Stability assessment models of power grid,
Real-time monitoring and assessment are carried out to the Transient Voltage Stability state of power grid.
Compared with prior art, the present invention have it is following the utility model has the advantages that
A kind of adjusting method that power grid Transient Voltage Stability sample set classification is unbalance proposed by the present invention, its advantage is that, phase
It is in power grid reality than the training sample in traditional method for obtaining training sample based entirely on time-domain-simulation, the method for the present invention
It is obtained on the basis of the method for operation of border, ensure that the reliability of training sample set from data source, for the history being originally taken
The classification unbalance of sample, by generating the pre- test sample of unstability under dispatching of power netwoks operational plan to the time-domain-simulation of forecast failure
This, and according to unstability forecast sample and stablize Euclidean distance relationship between historical sample to the qualification of unstability forecast sample into
Performing check is further ensured that the quality of unstability forecast sample, and unstability forecast sample collection is merged into historical sample collection and is used to dig
Classification learning process can be improved to mistake while effectively adjusting historical sample collection classification is unbalance in the training sample set for digging study
The bias of steady sample, enhancing power grid Transient Voltage Stability assessment models recall ability to unstability sample.
Detailed description of the invention
Fig. 1 is the embodiment power grid single line structural schematic diagram that the method for the present invention is related to；
Fig. 2 is the embodiment flow diagram of the method for the present invention；
Fig. 3 is to carry out the decision-tree model that classification learning obtains to training sample set in the method for the present invention.
Specific embodiment
In the following, being described further in conjunction with attached drawing and specific embodiment to the present invention:
Power grid single line structural schematic diagram involved in appraisal procedure of the present invention is as shown in Figure 1, middle part power grid shown in FIG. 1 is
An Application Example of the invention, the power grid are the heavy load in Fig. 1 by end regions, in total include 13 nodes, load is total
Amount is 6190 megawatts, and the present embodiment method flow is as shown in Figure 2, comprising the following steps:
(1) historical failure collection, node collection and the characteristic variable collection of power grid are obtained from the history log of power grid, and are received
Collect the characteristic variable value of each node in power grid under each historical failure and the operating status Z of power grid, power grid is in stable operation shape
State is denoted as Z=1, and power grid is in unstability operating status and is denoted as Z=-1, by the data acquisition system collected under a historical failure at one
Historical sample, total N0(the present embodiment is 720) a historical sample is integrated into a historical sample collection S0, difference statistical history sample
This concentration stablizes historical sample sum N1(the present embodiment is 665) and unstability historical sample sum N-1(it is 55 that this implementation, which is applied,
It is a), wherein N1+N-1=N0；
(2) node collection, the characteristic variable collection of present period power grid are obtained from the management and running platform of power grid, the following n is small
When (embodiment be 72 hours) in management and running plan and the forecast failure collection in n hours future, imitated using computer time domain
Various forecast failures progress N of the true method to the power grid in n hour following under management and running planpSecondary time-domain-simulation, point
The operation shape of each characteristic variable value and power grid of each node after meeting with failure in power grid during each time-domain-simulation is not recorded
State Z, wherein Z=1 represents power grid and is in steady operational status, and Z=-1 represents power grid and is in unstability operating status, sentences to Z
It is disconnected, if Z=-1, the one unstability forecast sample of Data Synthesis recorded during this time-domain-simulation is not made if Z=1
The synthesis of unstability forecast sample counts the unstability forecast sample number N of all synthesisp0(the present embodiment is 140), Np0A mistake
Steady forecast sample forms an initial predicted sample set Sp0；
(3) the historical sample collection S in step (1) is utilized0To the initial predicted sample set S in step (2)p0In all Np0It is a
Unstability forecast sample carries out approval test, and underproof sample is rejected, all through examining qualified unstability forecast sample shape
At qualified forecast sample collection Sp1, detailed process is as follows:
(3-1) utilizes the historical sample collection S in step (1)0With the initial predicted sample set S in step (2)p0Synthesis one
Test samples collection S1, wherein test samples sum is Nt=N0+Np0(860)；
The initial predicted sample set S of (3-2) any selecting step (2)p0In unstability forecast sample i, and from step (1)
Historical sample collection S0All N1It is arbitrarily chosen in a stable historical sample and stablizes historical sample j, calculated between the two samples
Euclidean distance d (i, j), wherein 1≤i≤Np0, 1≤j≤N1；
(3-3) successively calculates the test samples collection S of step (3-1)1In all NtIt is lost in a test samples and step (3-2)
Euclidean distance between steady forecast sample i unstability forecast sample i, all N that will be calculatedtThe maximum of a Euclidean distance numerical value
Value is denoted as d (i, Nt)；
(3-4) successively calculates test samples collection S in step (3-1)1All NtIt is steady in a test samples and step (3-2)
Determine the Euclidean distance between historical sample j, all N that will be calculatedtThe maximum value of a Euclidean distance numerical value is denoted as d (j, Nt)；
(3-5) is compared and judges to the obtained Euclidean distance of step (3-2)~(3-4), if d (i, Nt)≥d(i,j)
And d (j, Nt) >=d (i, j) then illustrates that unstability forecast sample i is unqualified, carries out step (3-7), if d (i, Nt) < d (i, j) and d
(j,Nt) >=d (i, j) or d (i, Nt) >=d (i, j) and d (j, Nt) < d (i, j) or d (i, Nt) < d (i, j) and d (j,
Nt) < d (i, j), then illustrate that unstability forecast sample i is qualified in this inspection, carries out step (3-6)；
(3-6) traversal history sample set S0In all N1A stable historical sample, repeat the above steps (3-2)~(3-
5) inspection result of unstability forecast sample i, is obtained；
(3-7) traverses initial predicted sample set Sp0In all Np0A unstability forecast sample, repeat the above steps (3-2)
~(3-6), obtains all Np0The inspection result of a unstability forecast sample；
All N that (3-8) read step (3-7) obtainsp0The inspection result of a unstability forecast sample, if unstability forecast sample
Inspection result be it is unqualified, then by the unstability forecast sample from initial predicted sample set Sp0Middle rejecting, if unstability forecast sample is examined
Result is tested for qualification, then by the unstability forecast sample in initial predicted sample set Sp0In retained, the unstability for counting qualified is pre-
Survey total sample number Np1(the present embodiment is 127), all Np1A qualified unstability forecast sample remained constitutes qualified prediction
Sample set Sp1；
(4) the qualified forecast sample collection S obtained using step (3)p1To adjust historical sample collection S in above-mentioned steps (1)0,
By Sp1It is merged into S0In, a training sample set is formed, to overcome classification unbalance, statistics shows unstability sample in history
It is respectively 7.63% and 21.49% that sample set and training sample, which concentrate proportion, thus illustrates incorporating unstability forecast sample
Afterwards, the classification unbalance of historical sample collection has obtained very big alleviation；
(5) classification learning is carried out using training sample set of the decision Tree algorithms to step (4), obtains a decision tree mould
Type, decision-tree model in one embodiment of the present of invention as shown in Fig. 2, in decision-tree model shown in Fig. 2 terminal node mark
Number 1 and -1 indicates the electric network state Z of output, and wherein Z=1 represents stabilization of power grids state, and Z=-1 represents power grid instability status, certainly
Internal node U_k indicates that the voltage characteristic variable of node k, P_k indicate the active power characteristic variable of node k in plan tree-model,
Q_k indicates the reactive power characteristic variable of node k, using cross validation mode to the classification accuracy P of decision-tree modelreWith call together
Return degree RecIt is tested, Pre=96.7%, Rec=96.2%, it is further contrast verification, by the historical sample collection of step (1)
It is directly regarded as training sample set and carries out classification learning, using point for the decision-tree model that cross validation mode obtains classification learning
Class accuracy rate PreWith degree of recalling RecIt is tested, Pre=92.2%, Rec=85.5%, thus illustrate unstability forecast sample to going through
The classification unbalance of history sample set has good regulating effect, not only increases whole classification accuracy, also significant to increase
Strong decision-tree model recalls ability to unstability sample, and decision-tree model shown in Fig. 2 is steady as the transient voltage of power grid
Determine assessment models, real-time monitoring and assessment are carried out to the Transient Voltage Stability state of power grid.
It follows that this method collects historical sample from power grid history log, by right under management and running plan
The time-domain-simulation of forecast failure generates unstability forecast sample, and the unstability forecast sample through inspection qualification is merged with historical sample,
It is adjusted to unbalance to the classification of historical sample collection, passes through the sample set being eased to classification unbalance
It practises, exports more structurally sound power grid Transient Voltage Stability assessment models.
Correspondingly, the present embodiment additionally provides a kind of tune that power grid Transient Voltage Stability sample set classification is unbalance
Section system, comprising:
Historical sample collection generation module is used to collect historical sample from power grid history log, be gone through with spanning set
History sample set Sp0；
Initial predicted sample set generation module is used for by raw to the time-domain-simulation of forecast failure under management and running plan
At unstability forecast sample, all unstability forecast samples form an initial predicted sample set Sp0；
Qualified forecast sample collection generation module passes through historical sample collection generation module historical sample collection S generatedp0It is right
Initial predicted sample set generation module initial predicted sample set S generatedp0In all unstability forecast samples carry out qualification
It examines, underproof unstability forecast sample is rejected, it is all through examining qualified unstability forecast sample to form qualified forecast sample
Collect Sp1；
Training sample set generation module passes through the qualified forecast sample obtained using qualified forecast sample collection generation module
Collect Sp1To adjust the historical sample collection S of above-mentioned steps initial predicted sample set generation module generation0, by Sp1It is merged into S0In, shape
At a training sample set；
Decision-tree model generation module, the training sample for using decision Tree algorithms to generate training sample set generation module
Collection carries out classification learning, obtains a decision-tree model, using decision-tree model as the Transient Voltage Stability assessment models of power grid,
Real-time monitoring and assessment are carried out to the Transient Voltage Stability state of power grid.
Since the working principle of above-mentioned modules and the process principle of the above method are identical, in the present embodiment just no longer
It repeats.
It will be apparent to those skilled in the art that can make various other according to the above description of the technical scheme and ideas
Corresponding change and deformation, and all these changes and deformation all should belong to the protection scope of the claims in the present invention
Within.
Claims (7)
1. a kind of adjusting method that power grid Transient Voltage Stability sample set classification is unbalance, which is characterized in that including
S1, historical sample collection S is collected from power grid history logp0；
S2, by generating unstability forecast sample, all unstability predictions to the time-domain-simulation of forecast failure under management and running plan
Sample forms an initial predicted sample set Sp0；
S3, the historical sample collection S in step S1 is utilizedp0To the initial predicted sample set S in step S2p0In all unstabilitys prediction
Sample carries out approval test, and underproof unstability forecast sample is rejected, all through examining qualified unstability forecast sample shape
At qualified forecast sample collection Sp1；
S4, the qualified forecast sample collection S obtained using step S3p1To adjust historical sample collection S in above-mentioned steps S10, by Sp1It closes
And arrive S0In, form a training sample set；
S5, classification learning is carried out using training sample set of the decision Tree algorithms to step S4, obtains a decision-tree model, it will certainly
Transient Voltage Stability assessment models of the plan tree-model as power grid, to the Transient Voltage Stability state of power grid carry out real-time monitoring and
Assessment.
2. the adjusting method that power grid Transient Voltage Stability sample set classification as described in claim 1 is unbalance, which is characterized in that institute
State step S1 specifically:
Historical failure collection, node collection and the characteristic variable collection of power grid are obtained from the history log of power grid, and are collected and respectively gone through
Under history failure in power grid the characteristic variable value of each node and power grid operating status Z, power grid is in steady operational status and is denoted as Z
=1, power grid is in unstability operating status and is denoted as Z=-1, by the data acquisition system collected under a historical failure at a history sample
This, total N0A historical sample is integrated into a historical sample collection S0, stablize historical sample sum N in statistical history sample set respectively1
With unstability historical sample sum N-1, wherein N1+N-1=N0。
3. the adjusting method that power grid Transient Voltage Stability sample set classification as claimed in claim 1 or 2 is unbalance, feature exist
In the step S2 specifically:
From the management and running platform of power grid obtain present period power grid node collection, characteristic variable collection, it is n hours following in tune
The forecast failure collection in operational plan and n hours future is spent, it is small in the following n to the power grid using computer time-domain simulation method
When interior management and running plan under various forecast failures carry out NpSecondary time-domain-simulation records electricity during each time-domain-simulation respectively
The operating status Z of each characteristic variable value and power grid of each node after meeting with failure in net, wherein Z=1 represents power grid and is in steady
Determine operating status, Z=-1 represents power grid and is in unstability operating status, judges Z, if Z=-1, by this time-domain-simulation
One unstability forecast sample of the Data Synthesis recorded in the process does not make the synthesis of unstability forecast sample if Z=1, and statistics is all
The unstability forecast sample number N of synthesisp0, Np0A unstability forecast sample forms an initial predicted sample set Sp0。
4. the adjusting method that power grid Transient Voltage Stability sample set classification as claimed in claim 3 is unbalance, which is characterized in that institute
Stating step S3 includes:
S31, the historical sample collection S in step S1 is utilized0With the initial predicted sample set S in step S2p0Synthesize an inspection sample
This collection S1, wherein test samples sum is Nt=N0+Np0；
The initial predicted sample set S of S32, any selecting step S2p0In unstability forecast sample i, and from the history sample of step S1
This collection S0All N1Arbitrarily chosen in a stable historical sample and stablize historical sample j, calculate Euclidean between the two samples away from
From d (i, j), wherein 1≤i≤Np0, 1≤j≤N1；
S33, the test samples collection S for successively calculating step S311In all NtA test samples and unstability forecast sample in step S32
Euclidean distance between i unstability forecast sample i, all N that will be calculatedtThe maximum value of a Euclidean distance numerical value be denoted as d (i,
Nt)；
S34, test samples collection S in step S31 is successively calculated1All NtStablize historical sample in a test samples and step S32
Euclidean distance between j, all N that will be calculatedtThe maximum value of a Euclidean distance numerical value is denoted as d (j, Nt)；
S35, the obtained Euclidean distance of step S32~S34 is compared and is judged, if d (i, Nt) >=d (i, j) and d (j, Nt)
>=d (i, j) then illustrates that unstability forecast sample i is unqualified, step S37 is carried out, if d (i, Nt) < d (i, j) and d (j, Nt)≥d
(i, j) or d (i, Nt) >=d (i, j) and d (j, Nt) < d (i, j) or d (i, Nt) < d (i, j) and d (j, Nt) < d (i,
J), then illustrate that unstability forecast sample i is qualified in this inspection, carry out step S36；
S36, traversal history sample set S0In all N1A stable historical sample, repeat the above steps S32~S35, obtains unstability
The inspection result of forecast sample i；
S37, traversal initial predicted sample set Sp0In all Np0A unstability forecast sample, repeat the above steps S32~S36, obtains
To all Np0The inspection result of a unstability forecast sample；
All N that S38, read step S37 are obtainedp0The inspection result of a unstability forecast sample, if unstability forecast sample examines knot
Fruit be it is unqualified, then by the unstability forecast sample from initial predicted sample set Sp0Middle rejecting, if unstability forecast sample inspection result
For qualification, then by the unstability forecast sample in initial predicted sample set Sp0In retained, count qualified unstability forecast sample
Total Np1, all Np1A qualified unstability forecast sample remained constitutes qualified forecast sample collection Sp1。
5. a kind of regulating system that power grid Transient Voltage Stability sample set classification is unbalance characterized by comprising
Historical sample collection generation module is used to collect historical sample from power grid history log, with spanning set history sample
This collection Sp0；
Initial predicted sample set generation module is used to lose by generating the time-domain-simulation of forecast failure under management and running plan
Steady forecast sample, all unstability forecast samples form an initial predicted sample set Sp0；
Qualified forecast sample collection generation module passes through historical sample collection generation module historical sample collection S generatedp0To initial
Forecast sample collection generation module initial predicted sample set S generatedp0In all unstability forecast samples carry out approval test,
Underproof unstability forecast sample is rejected, it is all through examining qualified unstability forecast sample to form qualified forecast sample collection Sp1；
Training sample set generation module passes through the qualified forecast sample collection S obtained using qualified forecast sample collection generation modulep1
To adjust the historical sample collection S of above-mentioned steps initial predicted sample set generation module generation0, by Sp1It is merged into S0In, form one
A training sample set；
Decision-tree model generation module, the training sample set for using decision Tree algorithms to generate training sample set generation module into
Row classification learning obtains a decision-tree model, using decision-tree model as the Transient Voltage Stability assessment models of power grid, to electricity
The Transient Voltage Stability state of net carries out real-time monitoring and assessment.
6. the regulating system that power grid Transient Voltage Stability sample set classification as claimed in claim 5 is unbalance, which is characterized in that institute
State historical sample collection generation module spanning set historical sample collection Sp0Detailed process are as follows:
Historical failure collection, node collection and the characteristic variable collection of power grid are obtained from the history log of power grid, and are collected and respectively gone through
Under history failure in power grid the characteristic variable value of each node and power grid operating status Z, power grid is in steady operational status and is denoted as Z
=1, power grid is in unstability operating status and is denoted as Z=-1, by the data acquisition system collected under a historical failure at a history sample
This, total N0A historical sample is integrated into a historical sample collection S0, stablize historical sample sum N in statistical history sample set respectively1
With unstability historical sample sum N-1, wherein N1+N-1=N0。
7. such as the unbalance adjusting method of power grid Transient Voltage Stability sample set classification described in claim 5 or 6, feature exists
In the initial predicted sample set generation module forms initial predicted sample set Sp0Detailed process are as follows:
From the management and running platform of power grid obtain present period power grid node collection, characteristic variable collection, it is n hours following in tune
The forecast failure collection in operational plan and n hours future is spent, it is small in the following n to the power grid using computer time-domain simulation method
When interior management and running plan under various forecast failures carry out NpSecondary time-domain-simulation records electricity during each time-domain-simulation respectively
The operating status Z of each characteristic variable value and power grid of each node after meeting with failure in net, wherein Z=1 represents power grid and is in steady
Determine operating status, Z=-1 represents power grid and is in unstability operating status, judges Z, if Z=-1, by this time-domain-simulation
One unstability forecast sample of the Data Synthesis recorded in the process does not make the synthesis of unstability forecast sample if Z=1, and statistics is all
The unstability forecast sample number N of synthesisp0, Np0A unstability forecast sample forms an initial predicted sample set Sp0。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810866566.0A CN108988347B (en) | 2018-08-01 | 2018-08-01 | Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810866566.0A CN108988347B (en) | 2018-08-01 | 2018-08-01 | Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108988347A true CN108988347A (en) | 2018-12-11 |
| CN108988347B CN108988347B (en) | 2020-09-22 |
Family
ID=64554847
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810866566.0A Active CN108988347B (en) | 2018-08-01 | 2018-08-01 | Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108988347B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110336270A (en) * | 2019-04-22 | 2019-10-15 | 清华大学 | A kind of update method of Power system transient stability prediction model |
| CN111244937A (en) * | 2020-01-09 | 2020-06-05 | 清华大学 | Method for screening serious faults of transient voltage stability of power system |
| CN111628501A (en) * | 2020-06-18 | 2020-09-04 | 国网山东省电力公司济南供电公司 | AC/DC large power grid transient voltage stability assessment method and system |
| CN114997063A (en) * | 2022-06-17 | 2022-09-02 | 华北电力大学 | Power grid transient stability prediction method and system based on cost sensitive support vector machine |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102074955A (en) * | 2011-01-20 | 2011-05-25 | 中国电力科学研究院 | Method based on knowledge discovery technology for stability assessment and control of electric system |
| CN104617574A (en) * | 2015-01-19 | 2015-05-13 | 清华大学 | Assessment method for transient voltage stabilization of load area of electrical power system |
| CN105139289A (en) * | 2015-09-06 | 2015-12-09 | 清华大学 | Power system transient state voltage stability evaluating method based on misclassification cost classified-learning |
| US20160239034A1 (en) * | 2013-10-22 | 2016-08-18 | Mehta Tech, Inc. | Methods and apparatus for detecting and correcting instabilities within a power distribution system |
| CN106849069A (en) * | 2017-03-13 | 2017-06-13 | 东北电力大学 | A kind of transient stability evaluation in power system method based on Pin SVM |
| JP6174271B2 (en) * | 2014-10-31 | 2017-08-02 | 株式会社日立製作所 | System stabilization control apparatus and method |
| CN107482621A (en) * | 2017-08-02 | 2017-12-15 | 清华大学 | A kind of Transient Voltage Stability in Electric Power System appraisal procedure based on voltage sequential track |
-
2018
- 2018-08-01 CN CN201810866566.0A patent/CN108988347B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102074955A (en) * | 2011-01-20 | 2011-05-25 | 中国电力科学研究院 | Method based on knowledge discovery technology for stability assessment and control of electric system |
| US20160239034A1 (en) * | 2013-10-22 | 2016-08-18 | Mehta Tech, Inc. | Methods and apparatus for detecting and correcting instabilities within a power distribution system |
| JP6174271B2 (en) * | 2014-10-31 | 2017-08-02 | 株式会社日立製作所 | System stabilization control apparatus and method |
| CN104617574A (en) * | 2015-01-19 | 2015-05-13 | 清华大学 | Assessment method for transient voltage stabilization of load area of electrical power system |
| CN105139289A (en) * | 2015-09-06 | 2015-12-09 | 清华大学 | Power system transient state voltage stability evaluating method based on misclassification cost classified-learning |
| CN106849069A (en) * | 2017-03-13 | 2017-06-13 | 东北电力大学 | A kind of transient stability evaluation in power system method based on Pin SVM |
| CN107482621A (en) * | 2017-08-02 | 2017-12-15 | 清华大学 | A kind of Transient Voltage Stability in Electric Power System appraisal procedure based on voltage sequential track |
Non-Patent Citations (4)
| Title |
|---|
| LIPENG ZHU，CHAO LU ET AL.: "Imbalance Learning Machine-Based Power System Short-Term Voltage Stability Assessment", 《 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS》 * |
| XINRAN ZHANG，CHAO LU ET AL.: "Stability Analysis and Controller Design of a Wide-Area Time-Delay System Based on the Expectation Model Method", 《IEEE TRANSACTIONS ON SMART GRID》 * |
| 朱利鹏，陆超等: "基于广域时序数据挖掘策略的暂态电压稳定评估", 《电网技术》 * |
| 朱利鹏，陆超等: "基于数据挖掘的区域暂态电压稳定评估", 《电网技术》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110336270A (en) * | 2019-04-22 | 2019-10-15 | 清华大学 | A kind of update method of Power system transient stability prediction model |
| CN110336270B (en) * | 2019-04-22 | 2021-02-02 | 清华大学 | Updating method of transient stability prediction model of power system |
| CN111244937A (en) * | 2020-01-09 | 2020-06-05 | 清华大学 | Method for screening serious faults of transient voltage stability of power system |
| CN111628501A (en) * | 2020-06-18 | 2020-09-04 | 国网山东省电力公司济南供电公司 | AC/DC large power grid transient voltage stability assessment method and system |
| CN114997063A (en) * | 2022-06-17 | 2022-09-02 | 华北电力大学 | Power grid transient stability prediction method and system based on cost sensitive support vector machine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108988347B (en) | 2020-09-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106154209A (en) | Electrical energy meter fault Forecasting Methodology based on decision Tree algorithms | |
| CN111008778A (en) | Method and system for diagnosing abnormity of metering points of transformer area | |
| CN105139289B (en) | A kind of power grid Transient Voltage Stability appraisal procedure for dividing cost classification learning based on mistake | |
| CN103886518A (en) | Early warning method for voltage sag based on electric energy quality data mining at monitoring point | |
| CN105117602A (en) | Metering apparatus operation state early warning method | |
| CN112183894B (en) | Gravity dam earthquake risk probability analysis method | |
| CN117554821A (en) | Energy storage battery on-line monitoring management method and system based on deep learning network | |
| CN108988347A (en) | A kind of adjusting method and system that power grid Transient Voltage Stability sample set classification is unbalance | |
| CN102623990B (en) | Low-frequency deloading parameter online checking method of self-adaptation power grid operation mode | |
| CN109376801A (en) | Blade of wind-driven generator icing diagnostic method based on integrated deep neural network | |
| CN105866725A (en) | Method for fault classification of smart electric meter based on cluster analysis and cloud model | |
| CN102545214A (en) | Method for evaluating reliability of wind farm having battery energy storage equipment | |
| CN112485747A (en) | Error state evaluation method for capacitor voltage transformer | |
| CN104407273B (en) | Electric energy quality disturbance source positioning method considering monitoring reliability | |
| CN105353304A (en) | Validation method of low voltage ride-through characteristic of electric model of wind turbine generator | |
| CN107437135A (en) | A kind of novel energy-storing selection method | |
| CN111950913A (en) | Micro-grid power quality comprehensive evaluation method based on node voltage sensitivity | |
| CN105930900A (en) | Method and system for predicting hybrid wind power generation | |
| CN107036808A (en) | Gearbox of wind turbine combined failure diagnostic method based on SVMs probability Estimation | |
| CN105139287A (en) | Comprehensive electric energy metering device state assessment method integrating safety domain | |
| CN111950131B (en) | Wind power plant output equivalent aggregation model construction method considering electricity limiting factors | |
| CN117077546B (en) | Power system load modeling method and system based on data driving | |
| CN105956755A (en) | Method and system for establishing quantitative relationship of general power line loss rate influencing factors | |
| CN109902133A (en) | Multi-source data correction processing method and system based on power grid arbitrary segmentation area | |
| CN103928923B (en) | A kind of network stationary power quality method for early warning based on sensitivity analysis |
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 |