CN108595884A - Power system transient stability appraisal procedure and device - Google Patents
Power system transient stability appraisal procedure and device Download PDFInfo
- Publication number
- CN108595884A CN108595884A CN201810439637.9A CN201810439637A CN108595884A CN 108595884 A CN108595884 A CN 108595884A CN 201810439637 A CN201810439637 A CN 201810439637A CN 108595884 A CN108595884 A CN 108595884A
- Authority
- CN
- China
- Prior art keywords
- point
- assessed
- sample
- assessment result
- feature vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
Abstract
A kind of power system transient stability appraisal procedure of offer of the embodiment of the present invention and device, the method includes：According to the key stato variable screened in advance, the feature vector of electric system point to be assessed is obtained；The mahalanobis distance between the feature vector of point to be assessed and the feature vector of each sample point in sample set is calculated separately, the sample point adjacent with point to be assessed is searched according to mahalanobis distance；According to categorised decision rule and the adjacent sample point, the assessment result of the point to be assessed is obtained.The embodiment of the present invention is by calculating the mahalanobis distance between point to be assessed and sample point feature vector, since mahalanobis distance has fully considered the correlation in feature vector between key stato variable, to make being accurately calculated for distance, the adjacent sample point for accurately searching point to be assessed is realized, the accuracy of assessment result is further increased.
Description
Technical field
The present embodiments relate to field of power, are assessed more particularly, to a kind of power system transient stability
Method and device.
Background technology
With the access of various new energy and the use of UHV transmission line, power grid becomes more complicated.Pass through
System is modeled in detail, that there are models is multiple for the method analyzed to electric system using conventional numeric computer sim- ulation method
Miscellaneous, the features such as speed is slower.And the use of the accumulation and big data method of operation of power networks data, to based on data and machine learning
The method predicted power system transient stability brings new thinking, can accelerate the speed of electric system simulation
Degree.And the prediction technique based on machine learning is not only analyzed if directly analyzed the dynamic data of all variables of power grid
Difficulty is larger, and is easy to neglect the information of key stato variable, reduces the accuracy of stability analysis.Therefore, usual feelings
Under condition, the dynamic data of electric system can be handled, the Partial Variable filtered out is classified or clustered, to electric power
The characteristic of system is studied.
And when being classified or being clustered, the distance between sample is a very important concept.Usual feelings
Under condition, during classification or cluster, the method that can use Euclidean distance weighs the distance of sample.But by
In before being classified or being clustered, corresponding processing and screening carried out to dynamic data, between the attribute of sample
There may be certain correlation.If directly using Euclidean distance, possibly can not count and these correlations, thus can not be accurate
Really portray the property of dynamic process of electrical power system.
Invention content
To solve the above-mentioned problems, the embodiment of the present invention provides one kind and overcoming the above problem or solve at least partly
State the power system transient stability appraisal procedure and device of problem.
According to a first aspect of the embodiments of the present invention, a kind of power system transient stability appraisal procedure, this method are provided
Including：According to the key stato variable screened in advance, the feature vector of electric system point to be assessed is obtained；It calculates separately to be assessed
Point feature vector and sample set in each sample point feature vector between mahalanobis distance, according to mahalanobis distance search with it is to be evaluated
Estimate a little adjacent sample point；According to categorised decision rule and the adjacent sample point, the assessment knot of the point to be assessed is obtained
Fruit.
Method provided in an embodiment of the present invention, by calculate the geneva between point to be assessed and sample point feature vector away from
From since mahalanobis distance has fully considered the correlation in feature vector between key stato variable, to make the calculating of distance
It is more accurate, the adjacent sample point for accurately searching point to be assessed is realized, the accuracy of assessment result is further increased.
Second aspect according to embodiments of the present invention provides a kind of power system transient stability apparatus for evaluating, the device
Including：First acquisition module, for according to the key stato variable that screens in advance, obtain the feature of electric system point to be assessed to
Amount；Searching module, in the feature vector and sample set for calculating separately point to be assessed between the feature vector of each sample point
Mahalanobis distance searches the sample point adjacent with point to be assessed according to mahalanobis distance；Second acquisition module, for according to categorised decision
The regular and described adjacent sample point obtains the assessment result of the point to be assessed.
According to a third aspect of the embodiments of the present invention, a kind of power system transient stability assessment equipment is provided, including：
At least one processor；And at least one processor being connect with processor communication, wherein：Memory, which is stored with, to be handled
The program instruction that device executes, the instruction of processor caller are able to carry out any in the various possible realization methods of first aspect
The power system transient stability appraisal procedure that the possible realization method of kind is provided.
According to a fourth aspect of the embodiments of the present invention, a kind of non-transient computer readable storage medium is provided, it is described non-
Transitory computer readable storage medium stores computer instruction, and computer instruction makes the various possibility of computer execution first aspect
Realization method in the power system transient stability appraisal procedure that is provided of any possible realization method.
It should be understood that above general description and following detailed description is exemplary and explanatory, it can not
Limit the embodiment of the present invention.
Description of the drawings
Fig. 1 is a kind of flow diagram of power system transient stability appraisal procedure of the embodiment of the present invention；
Fig. 2 is based under the uniform growth pattern for a kind of power system transient stability appraisal procedure of the embodiment of the present invention
The Transient Stability Evaluation accuracy schematic diagram of different distance；
Fig. 3 is a kind of non-homogeneous growth of load pattern of power system transient stability appraisal procedure of the embodiment of the present invention
Under the Transient Stability Evaluation accuracy schematic diagram based on different distance；
Fig. 4 is under a kind of uniform load growth pattern of power system transient stability appraisal procedure of the embodiment of the present invention
The Transient Stability Evaluation accuracy schematic diagram of algorithms of different；
Fig. 5 is under a kind of uniform load growth pattern of power system transient stability appraisal procedure of the embodiment of the present invention
Sensibility schematic diagram of the assessment algorithm of different distance to state variable number；
Fig. 6 is under a kind of uniform load growth pattern of power system transient stability appraisal procedure of the embodiment of the present invention
Robustness test result schematic diagram of the assessment algorithm of different distance to noise；
Fig. 7 is a kind of structural schematic diagram of power system transient stability apparatus for evaluating of the embodiment of the present invention；
Fig. 8 is a kind of structural schematic diagram of power system transient stability assessment equipment of the embodiment of the present invention.
Specific implementation mode
With reference to the accompanying drawings and examples, the specific implementation mode of the embodiment of the present invention is described in further detail.With
Lower embodiment is not limited to the range of the embodiment of the present invention for illustrating the embodiment of the present invention.
Transient stability evaluation in power system based on machine learning may include following steps：The selection of input data and feature
Extraction, sorting algorithm selection and for classification results analysis and that makes be correspondingly improved.Based on machine learning
Transient stability evaluation in power system, the principle mainly utilized is the method for " off-line training, in-circuit emulation ".Wherein, classifying
In algorithm, it is often necessary to calculate various distances, and classified according to distance, such as the distance of KNN neighbours, Kmeans distances and
Distance in similarity calculates.Distance usually has following fundamental characteristics：
D (x, x)=0//to the distance of oneself be 0；
d(x,y)>=0//apart from non-negative；
D (x, y)=d (y, x) // and apart from symmetry, the distance of the distance of A to B, B to A is equal；
d(x,z)+d(z,y)>=d (x, y) // Vector triangle, the sum of both sides are more than third side.
If can utilize time window as short as possible data quickly provide judgement as a result, so the distance is as most closed
Suitable, most effective distance.Under normal conditions, when carrying out Transient Stability Evaluation, the distance of use is Euclidean distance.
For said circumstances, as shown in Figure 1, the embodiment of the present invention provides a kind of power system transient stability assessment side
Method, including：101, according to the key stato variable screened in advance, the feature vector of electric system point to be assessed is obtained；102, divide
Do not calculate the mahalanobis distance between the feature vector of point to be assessed and the feature vector of each sample point in sample set, according to geneva away from
The sample point adjacent with point to be assessed from lookup；103, according to categorised decision rule and adjacent sample point, point to be assessed is obtained
Assessment result.
In a step 101, the key stato variable of electric system is the state variable after processing and screening, crucial shape
State variable can describe power system transient stability, and have correlation between key stato variable.Each point pair to be assessed
A moment is answered, at the time of point to be assessed is corresponding after the key stato variable of acquisition electric system, according to key stato variable
Obtain the feature vector of point to be assessed.
In a step 102, mahalanobis distance (Mahalanobis distance) is the covariance distance for indicating data.It is
A kind of method of effective similarity for calculating two unknown sample collection, unlike Euclidean distance, it considers various characteristics
Between contact (for example, when sample is the set of crowd, when being characterized in the height and weight of people, higher human body weight bigger) simultaneously
And be that scale is unrelated (scale-invariant), i.e., independently of measurement scale.It is μ for a mean value, covariance matrix is
The multivariable vector of ∑, mahalanobis distance are：
In formula, XiAnd XjRespectively two samples, S are covariance matrix.
For the evaluation point and sample point of electric system, due to the key state in evaluation point and the feature vector of sample point
Variable is chosen from same time series, so having correlation between key stato variable；Therefore, in order to away from
When from calculating, the correlation between key stato variable is fully considered, being measured using mahalanobis distance in the embodiment of the present invention
Point to be assessed is between each sample point in sample set at a distance from feature space.Therefore, the geneva of point and sample point to be assessed away from
From for：
In formula, D is mahalanobis distance, XiFor the feature vector of evaluation point, XjFor the feature vector of sample point, S is covariance square
Battle array.
It, can be true by the mahalanobis distance of point to be assessed and each sample point due to having multiple sample points in sample set
Make the sample point adjacent with point to be assessed.Wherein, adjacent sample point can be the sample that mahalanobis distance is less than setpoint distance
Point, or be ranked up according to the sequence of distance from small to large, using the sample point of preceding setting quantity as adjacent sample point.
In step 103, the assessment knot that KNN algorithms (also known as K nearest neighbour classifications algorithm) obtain point to be assessed can be based on
Fruit.Since sample point has corresponding classification, and classification can be stable or unstable.And point to be assessed is adjacent with sample point,
The type for indicating point to be assessed and the type of adjacent sample point have certain similitude.Therefore, categorised decision rule can be based on
And the classification of adjacent sample point treats evaluation point and classifies, and obtains the classification of point to be assessed, the category and assessment result phase
It is corresponding, for example, stablize or unstable.
Method provided in an embodiment of the present invention, by calculate the geneva between point to be assessed and sample point feature vector away from
From since mahalanobis distance has fully considered the correlation in feature vector between key stato variable, to make the calculating of distance
It is more accurate, the adjacent sample point for accurately searching point to be assessed is realized, the accuracy of assessment result is further increased.
Transient Stability Evaluation is carried out to electric system in the method for the dynamic data combination machine learning using electric system
When, although dynamic simulation data can include the information of more electric system to a certain extent, due to dynamic
Emulation data contain more POWER SYSTEM STATE variable, while also being emulated to longer time length, it is easy to make
The problem of at " dimension calamity ".For example, the state variable chosen has 2000, the time window length of emulation reaches 10s, and step-length is
0.01s, then emulate every time the result is that the matrix of a 2000*1000.When carrying out classifier training, if using
10000 samples, then easily lead to that sample is too big, it is can not being analyzed as a result, and time consumption for training it is longer, it is also unfavorable
In the multiple training parameter adjusting of grader and application on site.
Content based on above-mentioned principle and above-described embodiment, as a kind of alternative embodiment, the key stato variable is logical
Cross following manner screening：The multiple state variables for obtaining electric system simulation data respectively become each state using fft algorithm
Amount is handled；Feature extraction is carried out by multiple state variables to treated, filters out the key in multiple state variables
State variable.
Specifically, electric system simulation data can be dynamic waveform data.First using the methods of FFT to emulating number
According to time series data carry out feature extraction, then using feature extracting method filter out electric system key state become
Amount.Method provided in an embodiment of the present invention filters out key stato variable by feature extraction, reduces time consumption for training, favorably
In repeatedly trained parameter regulation and application on site.
Content based on above-described embodiment, as a kind of alternative embodiment, the pass filtered out in multiple state variables
After key-like state variable, further include：Dimensionality reduction is carried out to multiple key stato variables.Specifically, the drop of manifold learning can be used in dimensionality reduction
Dimension method or Principal Component Analysis, to reduce the dimension of key stato variable so that build sample in the training of grader later
The speed of this collection is accelerated.
Content based on above-described embodiment, it is described that multiple key stato variables are dropped as a kind of alternative embodiment
Further include after the step of dimension：According to the key stato variable, the sample set is built；Wherein, the sample set includes multiple
Sample point, it includes a feature vector that each sample point, which has corresponding classification, each sample point, and described eigenvector is by sample
The key stato variable at point corresponding moment obtains.
The classification of sample point is corresponding with the assessment result of point to be assessed, and classification can specifically include two classes：Stablize or not
Stablize.When building sample set, the classification of sample point specifically can be by being determined with lower boundary：
In formula, δmaxMaximum work angular difference after expression failure in electric system between arbitrary two generators.Based on above formula,
If η becomes negative in 15s, electric system i.e. can be considered unstable, i.e., the sample point at the moment be classified as it is unstable, it is no
It is then to stablize, to which the sample point to each moment in emulation data is all classified, classification collectively forms one with feature vector
Sample point.
Content based on above-described embodiment, as a kind of alternative embodiment, the categorised decision rule is advised for majority voting
Then.Majority voting rule is that the classification of point to be assessed is determined by the classification of adjacent sample point.For example, by calculating mahalanobis distance,
5 adjacent sample points are found altogether.In 5 adjacent sample points, the classifications of 4 sample points be it is unstable, 1 adjacent sample point
Classification is to stablize, due to 4>1, based on majority voting rule, confirm that the assessment result of point to be assessed is unstable.
Classification by being then based on adjacent sample point obtains the assessment result of point to be assessed, when using majority voting rule
When, in 5 adjacent sample points, the classification of 3 sample points is unstable, and the classification of 2 adjacent sample points is to stablize, due to 3>2,
It can judge that the assessment result of point to be assessed is unstable.But due to the sample point of unstable classification sample only more other than Stabilized
This point is 1 more, and the confidence level of the assessment result of point to be assessed is not high.Therefore, based in above-mentioned principle and above-described embodiment
Hold, as a kind of alternative embodiment, the step of the assessment result for obtaining point to be assessed after further include：Obtain the assessment knot
The confidence level of fruit, if judge know the assessment result confidence level be more than confidence threshold value, using the assessment result as
The transient stability evaluation in power system result；Otherwise, the assessment result for obtaining next evaluation point, until the confidence of assessment result
Degree is more than the confidence threshold value.
The quantity that calculating quantity of sample point corresponding to evaluation result of confidence level accounts for total adjacent sample point determines；
Such as in 5 adjacent sample points, the classification of 3 sample points is unstable, and the classification of 2 adjacent sample points is to stablize, then assesses
As a result it is 60% for unstable confidence level.Therefore, in online evaluation in real time, if the assessment result of current point to be assessed is set
Reliability is less than confidence threshold value, then continues to assess next evaluation point, until the confidence level of assessment result is higher than confidence level
Threshold value confirms that assessment result is effective, improves the validity of assessment assessment result.
Based on above-described embodiment, 10 machine of New England, 39 node system is used to generate the sample for Transient Stability Evaluation below
This set verifies the validity of proposition method.Three-phase ground short trouble is happened on 39 busbares, failure occur the moment be
Tf, the fault clearance moment is Tc.There are 10 groups of different fault clearance moment, are 1.01s, 1.02s ..., 1.1s respectively.Emulation
Step-length be 0.01s, and emulate continue for 15s.The time window of observation is 1.9s, wherein containing 1s before failure occurs
The data of 0.9s after data and failure occur.
The Transient Stability Evaluation result of each sample point is determined based on above formula (1).By to network load is horizontal and power grid
Topological structure is modified, and different sample sets can be obtained.In order to more fully to carried Transient Stability Evaluation method into
Row verification, can be used following kind of sample set.And it is much larger than unstable sample for stablizing sample, it is unfavorable for extraction and stablizes
The problem of with unstable feature, in order to improve training effectiveness, the method that uses Bagging integrated studies.
(1) network topology structure remains unchanged, and the node load level that number is 1 to 39 equal proportion from 0.9 to 1.1 becomes
Change, generates 11 groups of different Run-time scenarios.In conjunction with different faults type and trouble duration, 4290 groups of emulation can be obtained
As a result, wherein 3311 groups are to stablize example, 979 groups are unstable.3500 groups of examples are for training, and 790 groups for detecting.
(2) network topology structure remains unchanged, and the node load level that number is 1 to 10 equal proportion from 0.9 to 1.1 becomes
Change, the node load level that number is 11 to 20 remains unchanged, and generates 11 groups of different Run-time scenarios.In conjunction with different faults class
Type and trouble duration can obtain 4032 groups of simulation results, wherein 2150 groups are to stablize example, 1882 groups are unstable.
3500 groups of examples are for training, and 532 groups for detecting.
As shown in Fig. 2, accurate for the transient stability evaluation in power system based on different distance under uniform load growth pattern
Property.Abscissa is the length for the period chosen, and ordinate (AUC) is the accuracy rate of classification.As can be seen from Figure, uniform
Under load growth pattern, the grader based on mahalanobis distance has better classification performance, can utilize the shorter period, right
Stability of power system is assessed.
As shown in figure 3, accurate for the transient stability evaluation in power system based on different distance under non-uniform load growth pattern
True property.As can be seen from Figure, under non-uniform load growth pattern, the grader based on mahalanobis distance equally has better
Classification performance can utilize the shorter period, assess stability of power system.
As shown in figure 4, for the transient stability evaluation in power system accuracy of algorithms of different under uniform load growth pattern.
On data set under uniform load growth pattern, compare SVM algorithm of support vector machine, decision Tree algorithms, neural network algorithm
With performance of the KNN algorithms (method that i.e. the above embodiment of the present invention provides) under the data set based on mahalanobis distance.Pass through
Fig. 4 can be seen that in the example under uniform load growth pattern and non-uniform load growth pattern, based on mahalanobis distance
KNN graders are compared to SVM algorithm, decision Tree algorithms and neural network algorithm, it may have better classification performance.
As shown in figure 5, being sensitivity of the assessment algorithm to state variable number of different distance under uniform load growth pattern
Degree analysis.The data of 0.2s test grader after selecting system failure, change the number of state variable, can obtain Fig. 5.It can
To find that when using the KNN algorithms of mahalanobis distance, better classification performance can be reached using less state variable number.
As shown in fig. 6, the assessment algorithm for different distance under uniform load growth pattern tests knot to the robustness of noise
Fruit.White Gaussian noise is added in the data obtained under uniform load growth pattern, respectively to the algorithm based on various distances into
Row test, as a result as shown above.When using the KNN algorithms of mahalanobis distance, it can equally be reached more using the shorter time
Good classification performance.
Content based on above-described embodiment, an embodiment of the present invention provides a kind of assessments of power system transient stability to fill
It sets, which is used to execute the semantic technical ability creation method in above method embodiment.Ginseng
See Fig. 7, which includes：First acquisition module 701, for according to the key stato variable screened in advance, obtaining electric system
The feature vector of point to be assessed；Searching module 702, each sample in the feature vector and sample set for calculating separately point to be assessed
Mahalanobis distance between the feature vector of point searches the sample point adjacent with point to be assessed according to mahalanobis distance；Second obtains mould
Block 703, for according to categorised decision rule and the adjacent sample point, obtaining the assessment result of the point to be assessed.
Wherein, the key stato variable of electric system is the state variable after processing and screening, key stato variable
Power system transient stability can be described, and there is correlation between key stato variable.Each point to be assessed is one corresponding
At the moment, at the time of point to be assessed is corresponding after the key stato variable of acquisition electric system, the first acquisition module 701 is according to key
State variable obtains the feature vector of point to be assessed.
For the evaluation point and sample point of electric system, due to the key state in evaluation point and the feature vector of sample point
Variable is chosen from same time series, so having correlation between key stato variable；Therefore, in order to away from
When from calculating, the correlation between key stato variable is fully considered, being measured using mahalanobis distance in the embodiment of the present invention
Point to be assessed is between each sample point in sample set at a distance from feature space.Therefore, the geneva of point and sample point to be assessed away from
From for：
In formula, D is mahalanobis distance, XiFor the feature vector of evaluation point, XjFor the feature vector of sample point, S is covariance square
Battle array.
Due to having multiple sample points, searching module 702 that can pass through point to be assessed and each sample point in sample set
Mahalanobis distance determine the sample point adjacent with point to be assessed.Wherein, adjacent sample point can be less than for mahalanobis distance and set
The sample point of set a distance, or be ranked up according to the sequence of distance from small to large, using the sample point of preceding setting quantity as phase
Adjacent sample point.
Second acquisition module 703 can be based on the assessment that KNN algorithms (also known as K nearest neighbour classifications algorithm) obtain point to be assessed
As a result.Since sample point has corresponding classification, and classification can be stable or unstable.And point to be assessed and sample point phase
Neighbour, the type for indicating point to be assessed and the type of adjacent sample point have certain similitude.Therefore, the second acquisition module 703 can
Evaluation point is treated with the classification based on categorised decision rule and adjacent sample point to classify, and obtains the classification of point to be assessed,
The category is corresponding with assessment result, for example, stablizes or unstable.
Device provided in an embodiment of the present invention, by calculate the geneva between point to be assessed and sample point feature vector away from
From since mahalanobis distance has fully considered the correlation in feature vector between key stato variable, to make the calculating of distance
It is more accurate, the adjacent sample point for accurately searching point to be assessed is realized, the accuracy of assessment result is further increased.
Content based on above-described embodiment, as a kind of alternative embodiment, described device further includes：Judgment module is used for
The confidence level of the assessment result is obtained, the confidence level of the assessment result is more than confidence threshold value if judging to know, by institute
Assessment result is stated as the transient stability evaluation in power system result；Otherwise, the assessment result for obtaining next evaluation point, until
The confidence level of assessment result is more than the confidence threshold value.
Content based on above-described embodiment further includes screening module, for obtaining power train as a kind of alternative embodiment
Multiple state variables of system emulation data, are respectively handled each state variable using fft algorithm；By to treated
Multiple state variables carry out feature extraction, filter out the key stato variable in multiple state variables.
Content based on above-described embodiment, as a kind of alternative embodiment, screening module is additionally operable to multiple key states
Variable carries out dimensionality reduction.
Content based on above-described embodiment, as a kind of alternative embodiment, screening module includes construction unit, is used for basis
The key stato variable builds the sample set；Wherein, the sample set includes multiple sample points, and each sample point has
Corresponding classification, each sample point include a feature vector, and described eigenvector is corresponded to the key state at moment by sample point
Variable obtains.
An embodiment of the present invention provides a kind of power system transient stability assessment equipments, as shown in figure 8, the equipment packet
It includes：Processor (processor) 801, memory (memory) 802 and bus 803；Wherein, processor 801 and memory 802
Mutual communication is completed by bus 803 respectively；Processor 801 is used to call the program instruction in memory 802, to execute
The power system transient stability appraisal procedure that above-described embodiment is provided, such as including：According to the key state screened in advance
Variable obtains the feature vector of electric system point to be assessed；Calculate separately point to be assessed feature vector and sample set in various kinds
Mahalanobis distance between the feature vector of this point searches the sample point adjacent with point to be assessed according to mahalanobis distance；According to classification
Decision rule and the adjacent sample point obtain the assessment result of the point to be assessed.
The embodiment of the present invention also provides a kind of non-transient computer readable storage medium, the non-transient computer readable storage
Medium storing computer instructs, which makes computer execute the electric power system transient stability that corresponding embodiment is provided
Property appraisal procedure, such as including：According to the key stato variable screened in advance, obtain the feature of electric system point to be assessed to
Amount；The mahalanobis distance between the feature vector of point to be assessed and the feature vector of each sample point in sample set is calculated separately, according to
Mahalanobis distance searches the sample point adjacent with point to be assessed；According to categorised decision rule and the adjacent sample point, institute is obtained
State the assessment result of point to be assessed.
One of ordinary skill in the art will appreciate that：Realize that all or part of step of above method embodiment can pass through
The relevant hardware of program instruction is completed, and program above-mentioned can be stored in a computer read/write memory medium, the program
When being executed, step including the steps of the foregoing method embodiments is executed；And storage medium above-mentioned includes：ROM, RAM, magnetic disc or light
The various media that can store program code such as disk.
The embodiments such as power system transient stability assessment equipment described above are only schematical, wherein conduct
The unit that separating component illustrates may or may not be physically separated, the component shown as unit can be or
Person may not be physical unit, you can be located at a place, or may be distributed over multiple network units.It can root
According to actual need that some or all of module therein is selected to achieve the purpose of the solution of this embodiment.Ordinary skill
Personnel are not in the case where paying performing creative labour, you can to understand and implement.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
It is realized by the mode of software plus required general hardware platform, naturally it is also possible to pass through hardware.Based on this understanding, on
Stating technical solution, substantially the part that contributes to existing technology can be expressed in the form of software products in other words, should
Computer software product can store in a computer-readable storage medium, such as ROM/RAM, magnetic disc, CD, including several fingers
It enables and using so that a computer equipment (can be personal computer, server or the network equipment etc.) executes each implementation
Certain Part Methods of example or embodiment.
Finally it should be noted that：The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, it will be understood by those of ordinary skill in the art that：It still may be used
With technical scheme described in the above embodiments is modified or equivalent replacement of some of the technical features；
And these modifications or replacements, various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. a kind of power system transient stability appraisal procedure, which is characterized in that including：
According to the key stato variable screened in advance, the feature vector of electric system point to be assessed is obtained；
The mahalanobis distance between the feature vector of point to be assessed and the feature vector of each sample point in sample set is calculated separately, according to
Mahalanobis distance searches the sample point adjacent with point to be assessed；
According to categorised decision rule and the adjacent sample point, the assessment result of the point to be assessed is obtained.
2. according to the method described in claim 1, it is characterized in that, the key stato variable screens in the following manner：
The multiple state variables for obtaining electric system simulation data, are respectively handled each state variable using fft algorithm；
Feature extraction is carried out by multiple state variables to treated, the key state filtered out in multiple state variables becomes
Amount.
3. according to the method described in claim 2, it is characterized in that, the key state filtered out in multiple state variables becomes
After amount, further include：Dimensionality reduction is carried out to multiple key stato variables.
4. according to the method described in claim 3, it is characterized in that, after the progress dimensionality reduction to multiple key stato variables, go back
Including：
According to the key stato variable, the sample set is built；Wherein, the sample set includes multiple sample points, each sample
It includes a feature vector that this point, which has corresponding classification, each sample point, and described eigenvector corresponds to the moment by sample point
Key stato variable obtains.
5. according to the method described in claim 1, it is characterized in that, the categorised decision rule is majority voting rule.
6. method according to any one of claims 1 to 5, which is characterized in that the assessment result for obtaining point to be assessed
The step of after further include：
The confidence level of the assessment result is obtained, if judging to know, the confidence level of the assessment result is more than confidence threshold value,
Using the assessment result as the transient stability evaluation in power system result；Otherwise, the assessment result of next evaluation point is obtained,
Until the confidence level of assessment result is more than the confidence threshold value.
7. a kind of power system transient stability apparatus for evaluating, which is characterized in that including：
First acquisition module, for according to the key stato variable that screens in advance, obtain the feature of electric system point to be assessed to
Amount；
Searching module, in the feature vector and sample set for calculating separately point to be assessed between the feature vector of each sample point
Mahalanobis distance searches the sample point adjacent with point to be assessed according to mahalanobis distance；
Second acquisition module, for according to categorised decision rule and the adjacent sample point, obtaining commenting for the point to be assessed
Estimate result.
8. device according to claim 7, which is characterized in that further include：
Judgment module, the confidence level for obtaining the assessment result, if judging to know that the confidence level of the assessment result is more than
Confidence threshold value, then using the assessment result as the transient stability evaluation in power system result；Otherwise, next assessment is obtained
The assessment result of point, until the confidence level of assessment result is more than the confidence threshold value.
9. a kind of power system transient stability assessment equipment, which is characterized in that including：
At least one processor；
And at least one processor being connect with the processor communication, wherein：
The memory is stored with the program instruction that can be executed by the processor, and the processor calls described program to instruct energy
Enough methods executed as described in claim 1 to 6 is any.
10. a kind of non-transient computer readable storage medium, which is characterized in that the non-transient computer readable storage medium is deposited
Computer instruction is stored up, the computer instruction makes the computer execute the method as described in claim 1 to 6 is any.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810439637.9A CN108595884A (en) | 2018-05-09 | 2018-05-09 | Power system transient stability appraisal procedure and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810439637.9A CN108595884A (en) | 2018-05-09 | 2018-05-09 | Power system transient stability appraisal procedure and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108595884A true CN108595884A (en) | 2018-09-28 |
Family
ID=63636124
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810439637.9A Pending CN108595884A (en) | 2018-05-09 | 2018-05-09 | Power system transient stability appraisal procedure and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108595884A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829627A (en) * | 2019-01-04 | 2019-05-31 | 三峡大学 | A kind of safe confidence appraisal procedure of Electrical Power System Dynamic based on integrated study scheme |
| CN110224394A (en) * | 2019-05-29 | 2019-09-10 | 海南电网有限责任公司电力科学研究院 | Fourier decomposition algorithm suitable for non-stationary oscillation of power signal characteristic abstraction |
| CN110448306A (en) * | 2019-07-30 | 2019-11-15 | 东北大学 | A kind of online fault detection and diagnosis method based on continuous blood sugar monitoring system |
| CN111273126A (en) * | 2020-01-20 | 2020-06-12 | 湖南大学 | Power distribution network topology rapid sensing method |
| CN112594937A (en) * | 2020-12-16 | 2021-04-02 | 珠海格力电器股份有限公司 | Control method and device of water heater, electronic equipment and storage medium |
| CN112950009A (en) * | 2021-02-10 | 2021-06-11 | 北京淇瑀信息科技有限公司 | Resource quota allocation method and device and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160084919A1 (en) * | 2013-05-06 | 2016-03-24 | University Of Saskatchewan | Methods and apparatus for detection of transient instability and out-of-step conditions by state deviation |
| CN106503279A (en) * | 2015-09-06 | 2017-03-15 | 中国电力科学研究院 | A kind of modeling method for transient stability evaluation in power system |
| CN107171315A (en) * | 2017-05-18 | 2017-09-15 | 广东电网有限责任公司电力科学研究院 | A kind of transient stability evaluation in power system method based on RPTSVM |
-
2018
- 2018-05-09 CN CN201810439637.9A patent/CN108595884A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160084919A1 (en) * | 2013-05-06 | 2016-03-24 | University Of Saskatchewan | Methods and apparatus for detection of transient instability and out-of-step conditions by state deviation |
| CN106503279A (en) * | 2015-09-06 | 2017-03-15 | 中国电力科学研究院 | A kind of modeling method for transient stability evaluation in power system |
| CN107171315A (en) * | 2017-05-18 | 2017-09-15 | 广东电网有限责任公司电力科学研究院 | A kind of transient stability evaluation in power system method based on RPTSVM |
Non-Patent Citations (4)
| Title |
|---|
| 田芳等: "基于支持向量机综合分类模型和关键样本集的电力系统暂态稳定评估", 《电力系统保护与控制》 * |
| 肖先勇等: "电压暂降分类特征可比性、相关性及马氏距离分类法", 《中国电机工程学报》 * |
| 胡伟等: "基于深度学习的电力系统故障后暂态稳定评估研究", 《电网技术》 * |
| 许涛等: "基于统计学习理论的电力系统暂态稳定评估", 《中国电机工程学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829627A (en) * | 2019-01-04 | 2019-05-31 | 三峡大学 | A kind of safe confidence appraisal procedure of Electrical Power System Dynamic based on integrated study scheme |
| CN110224394A (en) * | 2019-05-29 | 2019-09-10 | 海南电网有限责任公司电力科学研究院 | Fourier decomposition algorithm suitable for non-stationary oscillation of power signal characteristic abstraction |
| CN110448306A (en) * | 2019-07-30 | 2019-11-15 | 东北大学 | A kind of online fault detection and diagnosis method based on continuous blood sugar monitoring system |
| CN111273126A (en) * | 2020-01-20 | 2020-06-12 | 湖南大学 | Power distribution network topology rapid sensing method |
| CN111273126B (en) * | 2020-01-20 | 2021-01-29 | 湖南大学 | Power distribution network topology rapid sensing method |
| CN112594937A (en) * | 2020-12-16 | 2021-04-02 | 珠海格力电器股份有限公司 | Control method and device of water heater, electronic equipment and storage medium |
| CN112594937B (en) * | 2020-12-16 | 2022-03-18 | 珠海格力电器股份有限公司 | Control method and device of water heater, electronic equipment and storage medium |
| CN112950009A (en) * | 2021-02-10 | 2021-06-11 | 北京淇瑀信息科技有限公司 | Resource quota allocation method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108595884A (en) | Power system transient stability appraisal procedure and device | |
| CN108520357B (en) | Method and device for judging line loss abnormality reason and server | |
| Roffo et al. | Infinite latent feature selection: A probabilistic latent graph-based ranking approach | |
| CN107103332B (en) | A kind of Method Using Relevance Vector Machine sorting technique towards large-scale dataset | |
| Zhang et al. | Apricot: A weight-adaptation approach to fixing deep learning models | |
| CN103716204A (en) | Abnormal intrusion detection ensemble learning method and apparatus based on Wiener process | |
| CN108090295A (en) | A kind of long-span cablestayed bridges Damages in Stay Cables recognition methods | |
| CN106503689A (en) | Neutral net local discharge signal mode identification method based on particle cluster algorithm | |
| CN112508442B (en) | Transient stability assessment method and system based on automatic and interpretable machine learning | |
| CN106503863A (en) | Based on the Forecasting Methodology of the age characteristicss of decision-tree model, system and terminal | |
| CN111523778A (en) | Power grid operation safety assessment method based on particle swarm algorithm and gradient lifting tree | |
| CN109118119A (en) | Air control model generating method and device | |
| CN108090615B (en) | Minimum frequency prediction method after power system fault based on cross entropy integrated learning | |
| Erturk et al. | Software fault prediction using Mamdani type fuzzy inference system | |
| CN108599152A (en) | The key stato variable choosing method and device of power system transient stability assessment | |
| CN108304567A (en) | High-tension transformer regime mode identifies and data classification method and system | |
| CN111680875A (en) | Unmanned aerial vehicle state risk fuzzy comprehensive evaluation method based on probability baseline model | |
| CN109840413A (en) | A kind of detection method for phishing site and device | |
| CN115600088A (en) | Distribution transformer fault diagnosis method based on vibration signals | |
| CN107977672A (en) | SF6 equipment secondary failure diagnostic methods based on mass data concurrent operation | |
| CN110705831A (en) | Power angle instability mode pre-judgment model construction method after power system fault and application thereof | |
| CN117156442A (en) | Cloud data security protection method and system based on 5G network | |
| Ntoutsi et al. | A general framework for estimating similarity of datasets and decision trees: exploring semantic similarity of decision trees | |
| CN116400168A (en) | Power grid fault diagnosis method and system based on depth feature clustering | |
| CN113506186B (en) | Method and device for screening disturbed track of power system, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180928 |
|
| RJ01 | Rejection of invention patent application after publication |