CN108876163A - The transient rotor angle stability fast evaluation method of comprehensive causality analysis and machine learning - Google Patents
The transient rotor angle stability fast evaluation method of comprehensive causality analysis and machine learning Download PDFInfo
- Publication number
- CN108876163A CN108876163A CN201810674177.8A CN201810674177A CN108876163A CN 108876163 A CN108876163 A CN 108876163A CN 201810674177 A CN201810674177 A CN 201810674177A CN 108876163 A CN108876163 A CN 108876163A
- Authority
- CN
- China
- Prior art keywords
- transient
- angle stability
- mode
- rotor angle
- cluster
- 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
- 230000001052 transient effect Effects 0.000 title claims abstract description 112
- 238000004458 analytical method Methods 0.000 title claims abstract description 49
- 238000011156 evaluation Methods 0.000 title claims abstract description 19
- 238000010801 machine learning Methods 0.000 title claims abstract description 19
- 238000000034 method Methods 0.000 claims abstract description 74
- 238000013136 deep learning model Methods 0.000 claims abstract description 20
- 238000004445 quantitative analysis Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 10
- 238000004088 simulation Methods 0.000 claims description 19
- 238000004364 calculation method Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000010248 power generation Methods 0.000 claims 1
- 230000006870 function Effects 0.000 abstract description 4
- 230000015572 biosynthetic process Effects 0.000 abstract description 2
- 239000000284 extract Substances 0.000 abstract 1
- 238000013473 artificial intelligence Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 201000004569 Blindness Diseases 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000007363 ring formation reaction Methods 0.000 description 1
- 230000006641 stabilisation Effects 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Classifications
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- 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
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses the transient rotor angle stability fast evaluation methods of a kind of comprehensive causality analysis and machine learning, based on safety on line analysis using the historical data of Transient Stability Evaluation function storage, history run mode sample is divided into several method of operation clusters, transient rotor angle stability nargin is described according to the difference between mode with linearization equations, and extracts power grid key feature amount;Using key feature amount as input quantity, cluster classifies situation as output quantity in a manner of the history run of formation, it constructs deep learning model and utilizes historical data sample training, contacting between current real time operation mode and history run mode is set up, estimates the Transient angle stability of current real time operation mode;The present invention can be while ensuring transient rotor angle stability precision of analysis, it is time-consuming to efficiently reduce calculating, the quantitative analysis of transient power angle stability of power grid quickly is obtained as a result, helping to find transient state operation risk present in power grid in time, promotes the safety operation level of power grid.
Description
Technical field
The present invention relates to power system security stability analysis technical field more particularly to a kind of comprehensive causality analysis and machines
The transient rotor angle stability fast evaluation method of study.
Background technique
In order to ensure the safe and stable operation of power grid especially extra-high voltage alternating current-direct current serial-parallel power grid, scheduling institutions at different levels need
The safe and stable operation situation of power grid can be quickly grasped, the risk point and degree of risk of operation of power networks are quickly positioned, to be
Risk prevention, the control decision of power grid provide foundation.The provincial and above dispatching of power netwoks control mechanism of most of the country at present
Powernet safety analysis application function is built, which is based on electric network model and real time operation mode data, to electricity
Running State carries out periodic analytical calculation, steady from the safety of many aspects comprehensive assessment power grid such as static state, transient state, dynamic
It is qualitative, and safety and stability problem or safety and stability hidden danger progress aid decision for discovery, conjunction is provided to management and running personnel
The regulation scheme suggestion of reason.
Transient stability analysis is the key function of safety on line analysis application, and core is whether analysis transient state generator rotor angle is lost
Surely, time-domain-simulation analytic approach or EEAC (the Extended Equal-Area on time-domain-simulation analysis foundation are mainly used at present
Criterion) quantitative analysis method is calculated based on stringent numerical value is carried out to electric network model and real time operation mode data, is obtained
The transient stability result of power grid out.The transient stability number of faults that usual provincial power network need to calculate is several hundred to thousands of, to protect
The calculating speed requirement for completing the whole network transient stability analysis in 5-10 minutes is demonstrate,proved, needs to dispose the meter of hundreds of CPU core numbers
Calculate resource.With the Quick Extended of Net Frame of Electric Network scale, all kinds of novel devices such as a large amount of wind-powered electricity generations, photovoltaic apparatus and UPFC add
Enter, computation complexity exponentially property ascendant trend, required computing resource or calculating time-consuming will also will be further increased.
At the same time, it is artificial to have ignited a new round for " the man-machine Great War " of Google AlphaGo and domestic and international go top expert
The upsurge of intelligence, it is various be reported in media and network about the report of artificial intelligence, comment and prospect, it allows it was recognized that artificial
The huge energy of the new technologies such as intelligence.Machine learning is a branch of artificial intelligence field, by going through computer largely
Information needed is excavated in history data, and therefrom learning law, and then intelligent recognition new samples or prediction future, to make computer
Correct response or judgement are made in the case where not programming clearly in advance.Machine learning is in autonomous driving vehicle, practical language
Sound identification, genome understanding etc. bring a large amount of helps.In field of power system, electric power system dispatching operation how is combined
Demand provides new solution for electric power system dispatching operation, analysis decision and thinks by new technologies such as big data and artificial intelligence
Road becomes the hot spot of Power System Analysis decision domain research.
The actual demand of electric system related scientific research, industry mechanism combination bulk power grid analysis decision, has carried out electric system
The correlative study of big data, artificial intelligence achieves certain achievement.Patent " the stability of power system based on historical data
The quick judgment method " (patent No.:2015110301902) in, by calculate the statistic of each quantity of state of power grid and electrical quantity with
The correlation of fault critical mute time obtains the characteristic quantity under each failure, then calculates current real-time mode and historical manner spy
Metric range between sign amount, and K- nearest neighbor algorithm is used according to metric range, obtain the degree of stability index of current way,
To judge the stability of current time power grid.A kind of patent " transient stability evaluation in power system based on online historical data
Method (the patent No.:2015108372153) in ", several static state amounts just are sifted out by artificial first, then to selected static shape
State amount is compressed, and power grid characteristic quantity is obtained, and on this basis, the extension and stabilization of unstability sample is carried out to historical data sample
The compression of sample forms and calculates sample, and carries out disaggregated model training and parameter optimization using SVM algorithm, forms disaggregated model,
The transient stability of power grid is assessed with this.The above method achieves certain effect, but does not fill to the application of historical analysis result
Point, using pure machine learning method, there are certain blindness, while by the method for artificial screening quantity of state, passing through to personnel
The dependence tested is larger.
Summary of the invention
The purpose of the present invention is to provide the transient rotor angle stability rapid evaluations of a kind of comprehensive causality analysis and machine learning
Method, realization fast and accurately differentiate the current real time operation mode Transient angle stability of power grid online, efficiently reduce temporarily
The calculating of state power-angle stability analysis is time-consuming, quickly obtain the quantitative analysis of transient power angle stability of power grid as a result, facilitate and
Transient stability operation risk present in Shi Faxian power grid, promotes the safety operation level of power grid.
To achieve the above object, the technical solution adopted by the present invention, it is specific as follows:
The transient rotor angle stability fast evaluation method of comprehensive causality analysis and machine learning, includes the following steps:
1) historical data sample is obtained;Safety on line based on EEAC quantitative analysis method analyzes application, obtains application fortune
The stored historical data of row, the data sample as analysis;Historical data includes history run mode data and operation side
The corresponding transient rotor angle stability simulation result of failure is respectively examined under formula;
2) history run mode clusters;In view of actual electric network operation has regularity and repeatability, power grid history is transported
Line mode data classify according to examination failure and safe and stable mode, judge transient state generator rotor angle caused by method of operation difference
Whether stability margin variation is applicable in linearization calculation formula, and history run mode is clustered corresponding method of operation cluster;
3) key feature amount selects;Based on history run mode data and transient rotor angle stability simulation result, analyze each
It examines under failure, influence of the operation of power networks quantity of state to Transient angle stability, selects the key feature amount under failure;
4) deep learning model construction and training:Using key feature amount as input quantity, using method of operation cluster as output
Amount, establishes deep learning model, using the key feature amount of selection and the corresponding method of operation cluster of each historical sample, to building
Deep learning model is trained, the deep learning model after being trained;
5) the affiliated method of operation cluster of real time operation mode calculates:The current real time operation mode input training of power grid is obtained
In deep learning model, obtain under each examination failure, method of operation cluster corresponding to current real time operation mode;
6) current real-time mode transient rotor angle stability nargin estimation:Choose the corresponding method of operation of current real time operation mode
The transient rotor angle stability simulation result of any history run mode data and history run mode in cluster, analysis are current real-time
Difference in the method for operation and the cluster of selection between history run mode is based on transient rotor angle stability nargin Method of fast estimating,
The transient rotor angle stability nargin for obtaining the current electric grid method of operation judges Transient angle stability according to angle stability nargin, and
Export Transient angle stability result;
7) current real-time mode simulation analysis:Based on EEAC quantitative analysis method, current real time operation mode is imitated
Very, seek Transient angle stability under current real-time mode as a result, include transient rotor angle stability nargin, divide group's mode and participate in because
Son;
8) simulation result sample process:By Transient angle stability under current real-time mode data and the real-time mode
As a result, being added in historical sample;And according to transient rotor angle stability Transient angle stability as a result, calculate under each examination failure when
Method of operation cluster belonging to preceding real-time mode judges that the method for operation cluster obtained show that mode cluster is with using deep learning model
It is no consistent, if inconsistent, re -training is carried out to deep learning model using historical sample;
9) 5) the power grid real time operation mode data for reacquiring a new round enter step, realize new round operation of power networks
The assessment of mode Transient angle stability.
History run mode described in step 2) clusters, and specifically includes following steps:
2-1) classify according to examination failure, safe and stable mode to each history run mode, it will be every in same class
A history run mode is considered as the same cluster;
2-2) judge between cluster and cluster because whether transient rotor angle stability nargin variation caused by method of operation difference is applicable in line
Property calculation formula, if be applicable in linearization calculation formula, two clusters are merged, same cluster is classified as;
2-3) repeat step 2-1) and 2-2), until completing the processing of all clusters under all classification.
Step 2-2) specifically include following steps:In identical examination failure FcUnder, history run mode SAIt is quantified to analyze
Transient rotor angle stability nargin out is ηA.c, history run mode SBThe quantified transient rotor angle stability nargin that obtains analyzed is
ηB.cIf the difference η of transient rotor angle stability narginA.c-ηB.cIt is poor to estimate with the transient rotor angle stability nargin as caused by method of operation difference
Different ηestBetween difference be less than setting threshold value ηline-set, i.e.,:|(ηA.c-ηB.c)-ηest| < ηline-set, then it is assumed that it is examining
Failure FcUnder, history run mode SAAnd SBTransient rotor angle stability nargin be applicable in linearization equations description.
Step 3) specifically includes following steps:
It 3-1) is based on EEAC quantitative analysis and is greater than setting threshold value λ as a result, choosing transient rotor angle stability and participating in the factorset-A's
Generator Status amount is as key feature amount;
3-2) calculate correlation I (X, the η of each history run mode quantity of state of power grid and Transient angle stabilityA), it chooses
Correlation is greater than setting threshold value Iset-AQuantity of state as key feature amount;
The power grid statistic closely related with examination failure 3-3) is selected as key feature amount, with the examination close phase of failure
The power grid statistic of pass includes power grid gross capability, total load and key sections power;
3-4) seek 3-1), 3-2) and the 3-3) union of key feature amount, the preceding N of screening wherein correlation maximumkeyA spy
Key feature amount of the sign amount as corresponding failure;NkeyIt is set according to power grid scale and calculated performance demand.
More preferably, step 1) is arranged a certain number of in the initial stage for lacking data sample in conjunction with operation of power networks feature
Power grid typical operation modes, and transient emulation analysis is carried out to typical operation modes using EEAC quantitative analysis method, formation is gone through
History data sample.
Operation of power networks quantity of state includes unit open state, unit output, power plant's power output, busbar voltage, phase angle, interconnection
Road trend, connecting transformer trend, load level, direct current are sent/are entered power, section power and route throwing outside and stop state.
Beneficial effect of the present invention includes:
The present invention discloses the transient rotor angle stability fast evaluation method of a kind of comprehensive causality analysis and machine learning, is based on
EEAC quantitative analysis, the historical data for making full use of safety on line analysis to store using Transient Stability Evaluation function (including power grid
History run mode data and corresponding transient stability result data), the comprehensive cause and effect based on strict mathematical model inference point
Analysis method and the machine learning method based on big data, realize under power grid real time operation mode Transient angle stability it is quick
Assessment.Due to avoiding the time-domain-simulation to electric network fault process, Transient Stability Evaluation can be substantially improved using the present invention
Calculating speed;Meanwhile by using the transient stability quantitative information in historical analysis result, the blind of pure machine learning is reduced
Mesh reduces the incidence relation of discovery contingency.The present invention can be used as the existing transient rotor angle stability based on Digital Simulation Analysis
Effective supplement of analysis method.
Detailed description of the invention
Present invention will be further explained below with reference to the attached drawings and examples;
Fig. 1 is a kind of transient rotor angle stability fast evaluation method process of comprehensive causality analysis and machine learning of the present invention
Figure.
Specific embodiment
The invention will be further described with reference to the accompanying drawing and by specific embodiment, and following embodiment is descriptive
, it is not restrictive, this does not limit the scope of protection of the present invention.
In order to make technological means of the invention, creation characteristic, workflow, application method reach purpose and effect, and it is
It is easy to understand the evaluation method with reference to specific embodiments the present invention is further explained.
The invention will be further described with reference to the accompanying drawings and in conjunction with example.
As shown in Figure 1, the transient rotor angle stability of this comprehensive causality analysis provided by the invention and machine learning is quickly commented
Estimate method, includes the following steps:
Step 1) obtains historical data sample;Safety on line based on EEAC quantitative analysis method analyzes application, and acquisition is answered
With running stored historical data, the data sample as analysis;Historical data includes history run mode data and fortune
The corresponding transient rotor angle stability simulation result of failure is respectively examined under line mode.In the initial stage for lacking enough data samples, knot
Operation of power networks feature, the representative power grid typical operation modes of setting certain amount are closed, and use the quantitative analysis side EEAC
Method carries out transient emulation analysis, history of forming data sample to typical operation modes;
Representative power grid typical operation modes should cover that the summer is big, the summer is small, the winter is big, the winter is small, maintenance, open loop, cyclization
Etc. different scenes, and by the way that different power generation-load level is arranged, it is ensured that the diversity of data sample.
Step 2)) history run mode clusters;In view of actual electric network operation has regularity and repeatability, power grid is gone through
History running mode data classifies according to examination failure and safe and stable mode, judges transient state caused by method of operation difference
Whether the variation of angle stability nargin is applicable in linearization calculation formula, and history run mode is clustered corresponding method of operation cluster.
There is consistent examination failure, identical safe and stable mode, and transient rotor angle stability nargin can in same method of operation cluster
By in a manner of difference be described with linearization equations.Same method of operation cluster may include multiple methods of operation and its analysis meter
It calculates as a result, may also only include 1 method of operation and its analysis result.
It include the following contents by history run model split to corresponding method of operation cluster:
2-1) classify according to examination failure, safe and stable mode to each history run mode, it will be every in same class
A history run mode is as 1 cluster;
2-2) because whether stability margin variation caused by method of operation difference is applicable in linearisation meter between analytic manifold A and cluster B
Formula description is calculated, if being applicable in, (cluster A and cluster B) is merged, and is classified as same cluster by the two clusters;Wherein, cluster A and cluster B, which refers to, appoints
It anticipates two method of operation clusters;
It 2-3) repeats the above steps, until completing the merging treatment of all classification, all clusters.
Wherein, whether the variation of stability margin caused by the difference according to the method for operation is applicable in the description of linearization calculation formula
Judgment method be:It is located at identical examination failure FcUnder, history run mode SAIt is quantified to analyze the transient rotor angle stability obtained
Nargin is ηA.c, history run mode SBThe quantified transient rotor angle stability nargin obtained of analyzing is ηB.cIf transient state generator rotor angle is steady
Determine the difference η of narginA.c-ηB.cWith pass through patent " the online transient safe and stable appraisal procedure based on forecast failure collection automatic screening "
(the patent No.:Formula (5) estimates difference by the transient rotor angle stability nargin of method of operation disparity estimation in 201710247481X)
ηestBetween difference be less than setting threshold value ηline-set, i.e.,:|(ηA.c-ηB.c)-ηest| < ηline-set, then it is assumed that in examination event
Hinder FcUnder, history run mode SAAnd SBTransient rotor angle stability nargin be applicable in linearization equations description.
The selection of step 3) key feature amount;Based on history run mode data and transient rotor angle stability simulation result, analysis
Under each examination failure, influence of the operation of power networks quantity of state to Transient angle stability selects the key feature amount under failure;
Selecting key feature amount includes the following contents:
It 3-1) is based on EEAC quantitative analysis and is greater than setting threshold value λ as a result, choosing transient rotor angle stability and participating in the factorset-A's
Generator Status amount is as key feature amount;
3-2) calculate correlation I (X, the η of each history run mode quantity of state of power grid and Transient angle stabilityA), it chooses
Correlation is greater than setting threshold value Iset-AQuantity of state as key feature amount;
3-3) artificial selected part and the power grid statistic for examining failure closely related, including power grid gross capability, total load
With key sections power, as key feature amount;
The union of above-mentioned three parts key feature amount 3-4) is sought, the preceding N of wherein correlation maximum is screenedkeyA characteristic quantity
Key feature amount as corresponding failure.NkeyIt is set according to power grid scale and calculated performance demand.
Operation of power networks quantity of state includes unit open state, unit output, power plant's power output, busbar voltage, phase angle, interconnection
Road trend, connecting transformer trend, load level, direct current are sent/are entered power, section power and route throwing outside and stop state.
Step 4) deep learning model construction and training:Using key feature amount as input quantity, using method of operation cluster as
Output quantity establishes deep learning model, using the key feature amount of selection and the corresponding method of operation cluster of each historical sample, to structure
The deep learning model built is trained, the deep learning model after being trained;
The affiliated method of operation cluster of step 5) real time operation mode calculates:The current real time operation mode input of power grid is trained
To deep learning model in, obtain under each examination failure, method of operation cluster corresponding to current real time operation mode;
The current real-time mode transient rotor angle stability nargin estimation of step 6):Choose the corresponding operation of current real time operation mode
The transient rotor angle stability simulation result of any history run mode data and history run mode in mode cluster, analysis are current
Difference in real time operation mode and the cluster of selection between history run mode " is based on forecast failure collection Automatic sieve using patent
The online transient safe and stable appraisal procedure of the choosing " (patent No.:Transient rotor angle stability nargin in 201710247481X) is quickly estimated
Calculation method obtains the transient rotor angle stability nargin of the current electric grid method of operation, judges that transient state generator rotor angle is steady according to angle stability nargin
It is qualitative, and export Transient angle stability result;
The current real-time mode simulation analysis of step 7):Based on EEAC quantitative analysis method, to current real time operation mode into
Row emulation seeks Transient angle stability under current real-time mode as a result, including transient rotor angle stability nargin, dividing group's mode and ginseng
With the factor;
Step 8) simulation result sample process:Transient state generator rotor angle under current real-time mode data and the real-time mode is steady
Qualitative results are added in historical sample;And according to transient rotor angle stability Transient angle stability as a result, calculating each examination failure
Method of operation cluster belonging to current real-time mode down judges the method for operation cluster obtained and obtains mode using deep learning model
Whether cluster is consistent, if inconsistent, carries out re -training to deep learning model using historical sample;
Step 9) reacquires the power grid real time operation mode data of a new round, enters step 5), realizes new round power grid
The assessment of method of operation Transient angle stability.
Those skilled in the art can to the present invention be modified or modification design but do not depart from think of of the invention
Think and range.Therefore, if these modifications and changes of the present invention belongs to the claims in the present invention and its equivalent technical scope
Within, then the present invention is also intended to include these modifications and variations.
Claims (6)
1. the transient rotor angle stability fast evaluation method of comprehensive causality analysis and machine learning, which is characterized in that including following step
Suddenly:
1) historical data sample is obtained;Safety on line based on EEAC quantitative analysis method analyzes application, obtains application operation institute
The historical data of storage, the data sample as analysis;Historical data includes under history run mode data and the method for operation
It is each to examine the corresponding transient rotor angle stability simulation result of failure;
2) history run mode clusters;In view of actual electric network operation has regularity and repeatability, to power grid history run side
Formula data classify according to examination failure and safe and stable mode, judge transient rotor angle stability caused by method of operation difference
Whether nargin variation is applicable in linearization calculation formula, and history run mode is clustered corresponding method of operation cluster;
3) key feature amount selects;Based on history run mode data and transient rotor angle stability simulation result, analyze in each examination
Under failure, influence of the operation of power networks quantity of state to Transient angle stability selects the key feature amount under failure;
4) deep learning model construction and training:Using key feature amount as input quantity, using method of operation cluster as output quantity, build
Vertical deep learning model, using the key feature amount of selection and the corresponding method of operation cluster of each historical sample, to the depth of building
Learning model is trained, the deep learning model after being trained;
5) the affiliated method of operation cluster of real time operation mode calculates:The depth that the current real time operation mode input training of power grid is obtained
In learning model, obtain under each examination failure, method of operation cluster corresponding to current real time operation mode;
6) current real-time mode transient rotor angle stability nargin estimation:It chooses in the corresponding method of operation cluster of current real time operation mode
The transient rotor angle stability simulation result of any history run mode data and history run mode, analyzes current real time execution
Difference in mode and the cluster of selection between history run mode is based on transient rotor angle stability nargin Method of fast estimating, obtains
The transient rotor angle stability nargin of the current electric grid method of operation judges Transient angle stability according to angle stability nargin, and exports
Transient angle stability result;
7) current real-time mode simulation analysis:Based on EEAC quantitative analysis method, current real time operation mode is emulated, is asked
Transient angle stability under current real-time mode is taken as a result, including transient rotor angle stability nargin, dividing group's mode and participate in the factor;
8) simulation result sample process:By Transient angle stability knot under current real-time mode data and the real-time mode
Fruit is added in historical sample;And according to transient rotor angle stability Transient angle stability as a result, calculating current under each examination failure
Method of operation cluster belonging to real-time mode judges the method for operation cluster obtained and whether obtains mode cluster using deep learning model
Unanimously, if it is inconsistent, re -training is carried out to deep learning model using historical sample;
9) 5) the power grid real time operation mode data for reacquiring a new round enter step, realize new round grid operation mode
The assessment of Transient angle stability.
2. the transient rotor angle stability fast evaluation method of comprehensive causality analysis according to claim 1 and machine learning,
It is characterized in that,
History run mode described in step 2) clusters, and specifically includes following steps:
2-1) classify according to examination failure, safe and stable mode to each history run mode, will be gone through each of in same class
The history method of operation is considered as the same cluster;
2-2) judge between cluster and cluster because whether transient rotor angle stability nargin variation caused by method of operation difference is applicable in linearisation
Two clusters are merged if being applicable in linearization calculation formula, are classified as same cluster by calculation formula;
2-3) repeat step 2-1) and 2-2), until completing the processing of all clusters under all classification.
3. the transient rotor angle stability fast evaluation method of comprehensive causality analysis according to claim 2 and machine learning,
It is characterized in that,
Step 2-2) specifically include following steps:In identical examination failure FcUnder, history run mode SAWhat quantified analysis obtained
Transient rotor angle stability nargin is ηA.c, history run mode SBThe quantified transient rotor angle stability nargin obtained of analyzing is ηB.cIf
The difference η of transient rotor angle stability narginA.c-ηB.cDifference η is estimated with the transient rotor angle stability nargin as caused by method of operation differenceest
Between difference be less than setting threshold value ηline-set, | (ηA.c-ηB.c)-ηest| < ηline-set, then it is assumed that in examination failure FcUnder,
History run mode SAAnd SBTransient rotor angle stability nargin be applicable in linearization equations description.
4. the transient rotor angle stability fast evaluation method of comprehensive causality analysis according to claim 1 and machine learning,
It is characterized in that,
The step 3) specifically includes following steps:
It 3-1) is based on EEAC quantitative analysis and is greater than setting threshold value λ as a result, choosing transient rotor angle stability and participating in the factorset-APower generation
Machine quantity of state is as key feature amount;
3-2) calculate correlation I (X, the η of each history run mode quantity of state of power grid and Transient angle stabilityA), choose correlation
Greater than setting threshold value Iset-AQuantity of state as key feature amount;
3-3) the selected power grid statistic closely related with examination failure is closely related with examination failure as key feature amount
Power grid statistic includes power grid gross capability, total load and key sections power;
3-4) seek 3-1), 3-2) and the 3-3) union of key feature amount, the preceding N of screening wherein correlation maximumkeyA characteristic quantity
Key feature amount as corresponding failure;NkeyIt is set according to power grid scale and calculated performance demand.
5. the transient rotor angle stability fast evaluation method of comprehensive causality analysis according to claim 1 and machine learning,
It is characterized in that,
A certain number of power grid typical case fortune are arranged in conjunction with operation of power networks feature in the initial stage for lacking data sample in step 1)
Line mode, and transient emulation analysis, history of forming data sample are carried out to typical operation modes using EEAC quantitative analysis method.
6. the transient rotor angle stability fast evaluation method of comprehensive causality analysis according to claim 1 and machine learning,
It is characterized in that,
Operation of power networks quantity of state includes unit open state, unit output, power plant's power output, busbar voltage, phase angle, interconnector tide
Stream, connecting transformer trend, load level, direct current are sent/are entered power, section power and route throwing outside and stops state.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810674177.8A CN108876163B (en) | 2018-06-27 | 2018-06-27 | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810674177.8A CN108876163B (en) | 2018-06-27 | 2018-06-27 | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108876163A true CN108876163A (en) | 2018-11-23 |
CN108876163B CN108876163B (en) | 2021-07-23 |
Family
ID=64294952
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810674177.8A Active CN108876163B (en) | 2018-06-27 | 2018-06-27 | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108876163B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109861206A (en) * | 2018-12-29 | 2019-06-07 | 国电南瑞科技股份有限公司 | A kind of transient rotor angle stability Contingency screening system and method based on support vector machines |
CN109902352A (en) * | 2019-01-24 | 2019-06-18 | 南瑞集团有限公司 | A kind of transient rotor angle stability analysis of key Characteristic Extraction method and system |
CN110163540A (en) * | 2019-06-28 | 2019-08-23 | 清华大学 | Electric power system transient stability prevention and control method and system |
CN110323743A (en) * | 2019-07-24 | 2019-10-11 | 国电南瑞科技股份有限公司 | A kind of clustering method and device of transient rotor angle stability evaluation history data |
CN110348540A (en) * | 2019-07-24 | 2019-10-18 | 国电南瑞科技股份有限公司 | Electrical power system transient angle stability Contingency screening method and device based on cluster |
CN110443724A (en) * | 2019-07-19 | 2019-11-12 | 河海大学 | A kind of electric system fast state estimation method based on deep learning |
CN110707695A (en) * | 2019-11-12 | 2020-01-17 | 国电南瑞科技股份有限公司 | Transient power angle stability margin calculation method and system based on artificial intelligence |
CN110969214A (en) * | 2019-12-18 | 2020-04-07 | 天津大学 | Transient state security domain online construction method based on support vector machine comprehensive model |
CN113114527A (en) * | 2021-03-15 | 2021-07-13 | 河池学院 | Micro-grid transient stability rapid discrimination system and method based on machine learning |
Citations (5)
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 |
WO2012122310A1 (en) * | 2011-03-08 | 2012-09-13 | Trilliant Networks, Inc. | System and method for managing load distribution across a power grid |
CN104578053A (en) * | 2015-01-09 | 2015-04-29 | 北京交通大学 | Power system transient stability prediction method based on disturbance voltage trajectory cluster features |
CN106504116A (en) * | 2016-10-31 | 2017-03-15 | 山东大学 | Based on the stability assessment method that operation of power networks is associated with transient stability margin index |
CN107800140A (en) * | 2017-10-18 | 2018-03-13 | 天津大学 | A kind of large user for considering load characteristic, which powers, accesses decision-making technique |
-
2018
- 2018-06-27 CN CN201810674177.8A patent/CN108876163B/en active Active
Patent Citations (5)
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 |
WO2012122310A1 (en) * | 2011-03-08 | 2012-09-13 | Trilliant Networks, Inc. | System and method for managing load distribution across a power grid |
CN104578053A (en) * | 2015-01-09 | 2015-04-29 | 北京交通大学 | Power system transient stability prediction method based on disturbance voltage trajectory cluster features |
CN106504116A (en) * | 2016-10-31 | 2017-03-15 | 山东大学 | Based on the stability assessment method that operation of power networks is associated with transient stability margin index |
CN107800140A (en) * | 2017-10-18 | 2018-03-13 | 天津大学 | A kind of large user for considering load characteristic, which powers, accesses decision-making technique |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109861206A (en) * | 2018-12-29 | 2019-06-07 | 国电南瑞科技股份有限公司 | A kind of transient rotor angle stability Contingency screening system and method based on support vector machines |
CN109861206B (en) * | 2018-12-29 | 2022-08-09 | 国电南瑞科技股份有限公司 | Transient power angle stability fault screening system and method based on support vector machine |
CN109902352A (en) * | 2019-01-24 | 2019-06-18 | 南瑞集团有限公司 | A kind of transient rotor angle stability analysis of key Characteristic Extraction method and system |
CN109902352B (en) * | 2019-01-24 | 2022-07-01 | 南瑞集团有限公司 | Extraction method and system for key characteristic quantity of transient power angle stability analysis |
CN110163540B (en) * | 2019-06-28 | 2021-06-15 | 清华大学 | Power system transient stability prevention control method and system |
CN110163540A (en) * | 2019-06-28 | 2019-08-23 | 清华大学 | Electric power system transient stability prevention and control method and system |
CN110443724B (en) * | 2019-07-19 | 2022-08-16 | 河海大学 | Electric power system rapid state estimation method based on deep learning |
CN110443724A (en) * | 2019-07-19 | 2019-11-12 | 河海大学 | A kind of electric system fast state estimation method based on deep learning |
CN110348540A (en) * | 2019-07-24 | 2019-10-18 | 国电南瑞科技股份有限公司 | Electrical power system transient angle stability Contingency screening method and device based on cluster |
CN110348540B (en) * | 2019-07-24 | 2021-06-01 | 国电南瑞科技股份有限公司 | Clustering-based method and device for screening transient power angle stability faults of power system |
CN110323743B (en) * | 2019-07-24 | 2020-08-25 | 国电南瑞科技股份有限公司 | Clustering method and device for transient power angle stability evaluation historical data |
CN110323743A (en) * | 2019-07-24 | 2019-10-11 | 国电南瑞科技股份有限公司 | A kind of clustering method and device of transient rotor angle stability evaluation history data |
CN110707695A (en) * | 2019-11-12 | 2020-01-17 | 国电南瑞科技股份有限公司 | Transient power angle stability margin calculation method and system based on artificial intelligence |
CN110969214A (en) * | 2019-12-18 | 2020-04-07 | 天津大学 | Transient state security domain online construction method based on support vector machine comprehensive model |
CN110969214B (en) * | 2019-12-18 | 2023-06-23 | 天津大学 | Transient security domain online construction method based on support vector machine comprehensive model |
CN113114527A (en) * | 2021-03-15 | 2021-07-13 | 河池学院 | Micro-grid transient stability rapid discrimination system and method based on machine learning |
CN113114527B (en) * | 2021-03-15 | 2022-11-18 | 河池学院 | Micro-grid transient stability rapid discrimination system and method based on machine learning |
Also Published As
Publication number | Publication date |
---|---|
CN108876163B (en) | 2021-07-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108876163A (en) | The transient rotor angle stability fast evaluation method of comprehensive causality analysis and machine learning | |
Liu et al. | A systematic approach for dynamic security assessment and the corresponding preventive control scheme based on decision trees | |
Liu et al. | Wind power plant prediction by using neural networks | |
Wang et al. | Decision tree based online stability assessment scheme for power systems with renewable generations | |
CN102934312B (en) | Energy production system and control thereof | |
CN103488869A (en) | Wind power generation short-term load forecast method of least squares support vector machine | |
CN104408562A (en) | Photovoltaic system generating efficiency comprehensive evaluation method based on BP (back propagation) neural network | |
CN108053128A (en) | A kind of Power Network Transient Stability fast evaluation method based on ELM and TF | |
CN109165604A (en) | The recognition methods of non-intrusion type load and its test macro based on coorinated training | |
CN106355308B (en) | A method of wind power integration system core equipment is recognized based on decision tree | |
CN102707256A (en) | Fault diagnosis method based on BP-Ada Boost nerve network for electric energy meter | |
CN107153845A (en) | A kind of isolated island detection method of the adaptive grid-connected photovoltaic system based on machine learning | |
CN106485594A (en) | A kind of main distribution integration incident response decision method | |
CN112632840A (en) | Power grid transient stability evaluation method based on adaptive differential evolution algorithm and ELM | |
CN110837915A (en) | Low-voltage load point prediction and probability prediction method for power system based on hybrid integrated deep learning | |
CN110705831A (en) | Power angle instability mode pre-judgment model construction method after power system fault and application thereof | |
CN104218571A (en) | Running state evaluation method for wind power generation equipment | |
CN115526258A (en) | Power system transient stability evaluation method based on Spearman correlation coefficient feature extraction | |
CN110783913A (en) | Group-based optimal power grid topology online optimization method considering expected accident set | |
CN114358116A (en) | Oil-immersed transformer fault diagnosis method and system and readable storage medium | |
Chelaru et al. | Decision support system to determine the replacement ranking of the aged transformers in electric distribution networks | |
Liu et al. | Dynamic security assessment of western Danish power system based on ensemble decision trees | |
CN114169395A (en) | Construction method and application of dominant instability mode recognition model of power system | |
CN113947502A (en) | Power transmission network planning evaluation method and device based on support vector machine | |
CN111967689A (en) | Model and method for wind power generation prediction by combining multivariate and stepwise linear regression and artificial neural network |
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 |