CN109193703A - Consider the electric power system transient stability key feature selection method of classification lack of uniformity - Google Patents
Consider the electric power system transient stability key feature selection method of classification lack of uniformity Download PDFInfo
- Publication number
- CN109193703A CN109193703A CN201811301951.7A CN201811301951A CN109193703A CN 109193703 A CN109193703 A CN 109193703A CN 201811301951 A CN201811301951 A CN 201811301951A CN 109193703 A CN109193703 A CN 109193703A
- Authority
- CN
- China
- Prior art keywords
- feature
- transient stability
- sample
- accuracy rate
- key feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/24—Arrangements for preventing or reducing oscillations of power in networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/26—Arrangements for eliminating or reducing asymmetry in polyphase networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/50—Arrangements for eliminating or reducing asymmetry in polyphase networks
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Consider the electric power system transient stability key feature selection method of classification lack of uniformity, comprising the following steps: S1, the primitive character collection that Transient Stability Evaluation is constructed using the time dimension statistic of electric state;S2, the sample set for generating preset quantity;S3, comprehensive evaluation index class weights accuracy rate CWA is defined according to Transient Stability Evaluation feature selecting interpretational criteria;S4, Transient Stability Evaluation key feature is screened using the package method that sweep forward and the learning machine that transfinites combine.The present invention proposes two criterion for needing to meet for key feature selection in the unbalanced characteristic for being directed to electric power system transient stability, and on this basis, proposes the evaluation index that class weights accuracy rate index is selected as key feature.The screening criteria characterized by new evaluation index, it is searched for using the key feature that the package method that sweep forward and the learning machine that transfinites combine carries out transient stability, excavating has the higher key feature for debating knowledge and magnanimity to unstability class, meets the actual demand of electricity net safety stable analysis.
Description
Technical field
The present invention relates to power system security stability analysis technical fields, and in particular, to considers classification lack of uniformity
Electric power system transient stability key feature selection method.
Background technique
In recent years, with the fast development of the scale application of Wide Area Measurement System and computer technology, it is based on artificial intelligence
The transient stability evaluation in power system of energy technology receives the extensive concern of scholars.Generally, based on artificial intelligence technology
Transient Stability Evaluation is taken as the classification problem of a two-mode to handle, that is, is divided into stable and two class of unstability.Input feature vector is
An important factor for influencing artificial intelligence disaggregated model performance, the existing feature set applied to transient stability classification are mostly according to expert
The problems such as experience carries out artificial selection, and there are redundancies weakens disaggregated model performance.Therefore, the pass of Transient Stability Evaluation is studied
Key feature selection approach is of great significance.
For the feature selection issues of Transient Stability Prediction, existing feature selection approach can be divided into two classes, and the first kind was
Filter formula method: it is sorted according to significance level of the feature to classification criterion, taking the forward feature that sorts is input feature vector;Second class is
Embedded methods: feature selecting is embedded in classifier training, is assessed the classifying quality of character subset and then is preferably inputted spy
Sign.The evaluation index of feature selecting is an important factor for influencing input feature vector selection result, to be currently used for transient stability classification
Feature selection approach is selected mostly using classification accuracy as criterion so that the higher key feature of accuracy rate is as the defeated of disaggregated model
Enter feature.
In fact, traditional classification accuracy is not appropriate for conduct for electric power system transient stability classification problem
The evaluation index of feature selecting has ignored the unbalanced characteristic of transient stability classification problem.Due to unstability event occur it is general
Rate is lower, stablizes sample and unstability sample proportion and unbalanced in sample set, unstability sample size is less than even far fewer than steady
Determine sample size, according to classification accuracy as evaluation criterion, the phenomenon that being easy to appear " big number eat decimal ", leads to overall point
The problem that class accuracy rate is higher but the accuracy rate of unstability sample is lower.It illustrates, it is assumed that include 95% sample in sample set
For Stabilized, all samples are all classified as to stablize sample under extreme case, classification accuracy may be up to 95%, but due to unstability
Sample all classification mistake, assessment result lose practical significance.Therefore, because existing transient stability feature selection approach does not consider
The lack of uniformity of Transient Stability Evaluation, it is difficult to meet the actual demand of electricity net safety stable analysis.
Summary of the invention
The object of the invention is that the shortcomings that overcoming the above-mentioned prior art and deficiency, provide and consider classification lack of uniformity
Electric power system transient stability key feature selection method, this method are mentioned in the unbalanced characteristic for being directed to electric power system transient stability
Two criterion for needing to meet for key feature selection out, and on this basis, propose the index conduct of class weights accuracy rate
The evaluation index of key feature selection.The screening criteria characterized by new evaluation index utilizes sweep forward and the learning machine that transfinites
The package method combined carries out the key feature search of transient stability, and excavating has the higher key for debating knowledge and magnanimity special unstability class
Sign.
Technical solution used by the present invention solves the above problems is:
Consider the electric power system transient stability key feature selection method of classification lack of uniformity, comprising the following steps:
S1, the primitive character collection that Transient Stability Evaluation is constructed using the time dimension statistic of electric state;
Feature and stable state before and after S2, integration system failure, and generate the sample set of preset quantity;
S3, comprehensive evaluation index class weights accuracy rate is defined according to Transient Stability Evaluation feature selecting interpretational criteria
CWA;
S4, firstly, firstly, sample set is randomly divided into training set and test set, then, primitive character is concentrated single special
The classification capacity of sign is assessed in training set, and assesses feature using the method that sweep forward and the learning machine that transfinites combine
The classification performance of subset, and using the maximum character subset of class weights accuracy rate CWA as essential signature sets, finally, using surveying
The data of examination collection are tested for the property essential signature sets.
Further, primitive character collection described in step S1 includes the flow state and event before electric power system fault occurs
Electrical response curve after barrier generation, wherein flow state feature mainly includes the load level of system, generator output level
With node voltage level;Electrical response feature includes that instantaneous and the generator amature angle at failure removal moment, rotor occur for failure
The condition responsives features such as angular speed, rotor angular acceleration, active power impact and rotor kinetic energy.
Further, step S2 further includes following sub-step:
S201, on the basis of benchmark trend, using monte carlo method stochastical sampling generate system load level size,
Active power output is adjusted according to generator capacity size equal proportion, executes Load flow calculation and calculation of tidal current is checked,
If trend does not restrain or certain electrical variables are out-of-limit, the corresponding method of operation is not received, on the contrary then receive the method for operation;
S202, under the feasible method of operation, it is random to set fault disturbance scene, utilize time-domain simulation method to carry out transient state
Stability analysis, obtains the transient response track of each generator, and judges the stable state of system after failure generation;
Feature and stable state before and after S203, integration system failure, form transient stability sample;
S204, step S201 to step S203 is repeated, until generating the sample set of preset quantity, and sample set is returned
One change processing.
Further, Transient Stability Evaluation feature selecting interpretational criteria described in step S3: first is that overall evaluation criterion
Should with unstability sample and stablize that the relative populations of sample are unrelated, avoid the influence of unbalanced data volume;Second is that unstability sample should be protruded
This importance selects the feature for having stronger identification capability to unstability sample.
Further, comprehensive evaluation index class weights accuracy rate CWA described in step S3 are as follows:
In formula: TSR and TUR is respectively the accuracy rate of Stabilized sample and unstability class sample;α is adjustable factors, is used to table
Levy the significance level of Stabilized and unstability class;Wherein, the comprehensive evaluation index class weights accuracy rate index meets transient state
Stability Assessment feature selecting interpretational criteria.
Further, step S4 further includes following sub-step:
S401, sample set is randomly divided into training set and test set, carries out key feature selection and classification using training set
Device training;
S402, it is scanned for using sweep forward method, defined feature subset is empty set first, in training set, is utilized
The learning machine that transfinites concentrates the classification capacity of single feature to assess primitive character, according to the size of class weights accuracy rate value
It sorts to single feature, selects class weights accuracy rate to be worth maximum feature and character subset is added;
S403, remaining feature are combined with the feature in character subset one by one, are classified using learning machine is transfinited to assemblage characteristic
Performance Evaluation selects class weights accuracy rate value maximum according to the class weights accuracy rate of assemblage characteristic to remaining feature ordering
Feature be added character subset;
S404, step S403 is repeated, until character subset has been added in all features, selection corresponds to class in search process
Not Jia Quan the maximum character subset of accuracy rate be essential signature sets;
S405, the essential signature sets obtained according to screening, are tested for the property using the data of test set.
To sum up, the beneficial effects of the present invention are:
The present invention proposes to be applied to key feature selection method on the basis of analysis transient stability classification unbalanced characteristic
Two criterion, define evaluation index of the unbalanced accuracy rate index of classification as feature selecting on this basis.Not with classification
Balanced accuracy rate is target, carries out key feature search using sweep forward method and the learning machine algorithm that transfinites, excavates to mistake
Steady class event has the key feature of stronger identification capability, meets the actual demand of electricity net safety stable analysis.
Detailed description of the invention
Fig. 1 is sample set product process figure of the invention.
Fig. 2 is traditional accuracy rate index and the unbalanced accuracy rate index comparison diagram of classification.
Fig. 3 is sweep forward schematic diagram of the present invention.
Fig. 4 is transient state stable key feature selecting flow chart of the present invention.
Fig. 5 is flow diagram of the invention.
Specific embodiment
In order to solve applied to the feature set of transient stability classification to be carried out manually according to expertise in the prior art
The case where the problems such as selection, there are redundancies, reduction disaggregated model performance, the present invention is in the analysis unbalanced spy of transient stability classification
Property on the basis of, propose two criterion for being applied to key feature selection method, it is unbalanced accurate to define classification on this basis
Evaluation index of the rate index as feature selecting.Using the unbalanced accuracy rate of classification as target, using sweep forward method and transfinite
Learning machine algorithm carries out key feature search, excavates the key feature for having stronger identification capability to unstability class event, meets
The actual demand of electricity net safety stable analysis.The present invention will now be described in further detail with reference to the accompanying drawings and the accompanying drawings,
Embodiments of the present invention are not limited thereto, and the example that only present invention applies in figure does not have the principle of the present invention
Essential constraint.
Embodiment:
As shown in Figure 1, Figure 2, shown in Fig. 3, Fig. 4 and Fig. 5, consider the electric power system transient stability key feature of classification lack of uniformity
Selection method;
The building of primitive character collection;
Utilize the primitive character collection of the time dimension statistic construction Transient Stability Evaluation of electric state, including electric system
The electrical response curve after flow state and failure generation before failure generation;Flow state feature mainly includes the load of system
Horizontal, generator output level and node voltage are horizontal;
Electrical response feature is divided into failure and temporal characteristics and failure removal moment feature occurs, and instantaneous, power generation occurs for failure
Machine is impacted by unbalanced power, and temporal characteristics occur as failure using generator acceleration and active power impact;Failure
The moment is cut off, using shapes such as generator amature angle, rotor velocity, rotor angular acceleration, active power impact and rotor kinetic energy
State response is used as feature;
Flow state feature before resultant fault and the electrical response feature after failure, form the original of Transient Stability Evaluation
Feature set.
1 primitive character collection of table
Sample set generates;
On the basis of benchmark trend, the load level size of system is generated using monte carlo method stochastical sampling, according to
Active power output is adjusted to generator capacity size equal proportion, Load flow calculation is executed and calculation of tidal current is checked, if damp
Stream is not restrained or certain electrical variables are out-of-limit, then the corresponding method of operation is not received, on the contrary then receive the method for operation;
It is random to set fault disturbance scene under the feasible method of operation, transient stability is carried out using time-domain simulation method
Analysis, obtains the transient response track of each generator, and judges the stable state of system after failure generation, before integration system failure
Feature and stable state afterwards form sample;
It repeats the above steps, the sample set until generating preset quantity;The flow chart that sample set generates is as shown in Figure 1.
Transient Stability Evaluation feature selecting interpretational criteria and index;
Unstability sample and the lack of uniformity for stablizing sample are the key properties of electric power system transient stability classification problem, mainly
Embody both ways: first is that unstability sample and the sample size for stablizing sample are unbalanced, the quantity of unstability sample is remote less than even
Less than the quantity for stablizing sample;Second is that unstability sample and the significance level for stablizing sample are unbalanced, the significance level of unstability sample
Much larger than stablizing sample;
For the key feature select permeability of Transient Stability Evaluation, following two criterion: criterion one: overall merit should be met
Criterion should with unstability sample and stablize that the relative populations of sample are unrelated, avoid the influence of unbalanced data volume;Criterion two: it should protrude
The importance of unstability sample selects the feature for having stronger identification capability to unstability sample;
Generally, the classification performance of transient stability status predication can be indicated with confusion matrix, such as following table;
2 confusion matrix of table
Wherein, the element in matrix is to stablize sample to be classified as stable number TS respectively, stablizes sample and is divided by mistake
The number FU of unstability, unstability sample are divided into stable number FS by mistake, and unstability sample is divided into the number TU of unstability.
According to definition, stablizing sample accuracy rate TSR and unstability sample accuracy rate TUR can be respectively indicated are as follows:
According to above-mentioned criterion, comprehensive evaluation index class weights accuracy rate CWA is defined are as follows:
In formula: α is adjustable factors, for characterizing the significance level of Stabilized and unstability class.
From the above equation, we can see that CWA index only with stablize sample and the respective accuracy of unstability sample is related, with two class samples
Relative populations are unrelated, avoid the influence of unbalanced data volume, meet the criterion one of feature selecting;In addition, being added in CWA index
Adjustable factors α is used to characterize the significance level of Stabilized and unstability class;Wherein, α=1 indicates the significance level phase of TSR and TUR
Together, α>1 indicates that TUR has bigger influence, and α<1 indicates that TSR has bigger influence.
It, can due to the important significance level greater than Stabilized of unstability class for transient stability status predication
The value range of factor-alpha is adjusted to be greater than 1;The addition of adjustable factors enables CWA to protrude the significance level of unstability class, and adjustable factors are got over
Greatly, show to consider that the specific gravity of unstability class is bigger in overall performance, there is stronger identification energy to unstability class so as to filter out
The key feature of power meets the criterion two of feature selecting.
By traditional accuracy rate Acc index definition it is found that its expression formula are as follows:
In formula: Ns, Nu and N are respectively to stablize sample number, unstability sample number and total number of samples.From the above equation, we can see that traditional
Accuracy rate can regard the linear weighted function of TSR and TUR as, and weight factor is related with the relative populations for stablizing sample and unstability sample.
By being more than even it is found that stablizing sample size in sample set far more than unstability sample the characteristics of electric power system transient stability
This quantity, it is assumed that stablizing sample and unstability sample size ratio is 3:1 and 6:1, and adjustable factors take 3 and 6 respectively, with contour picture
The relational graph of TSR, TUR and Acc and CWA, comparison diagram are as shown in Figure 2 out.
Transient Stability Evaluation key feature is screened using the package method that sweep forward and the learning machine that transfinites combine;Including with
Lower sub-step:
S401, to sample set normalized, be randomly divided into training set and test set, utilize training set carry out key feature
Selection and classifier training;
S402, it is scanned for using sweep forward method, defined feature subset is empty set first, in training set, is used
The learning machine that transfinites concentrates the classification capacity of single feature to assess primitive character, according to the size of class weights accuracy rate value
It sorts to single feature, selects class weights accuracy rate to be worth maximum feature and character subset is added;The schematic diagram of sweep forward is such as
Shown in Fig. 3;
S403, remaining feature are combined with the feature in character subset one by one, according to the class weights accuracy rate of assemblage characteristic
To remaining feature ordering, selects class weights accuracy rate to be worth maximum feature and character subset is added;
S404, step S403 is repeated, until character subset has been added in all features, selection corresponds to class in search process
Not Jia Quan the maximum character subset of accuracy rate be essential signature sets;
S405, the essential signature sets obtained according to screening, are tested for the property using the data of test set;Transient stability is commented
The basic procedure for estimating key feature selection is as shown in Figure 4.
In short, the electric power system transient stability key feature selection method of the considerations of proposing classification lack of uniformity, is being analyzed
On the basis of the unbalanced characteristic of transient stability classification, two criterion for being applied to key feature selection method are proposed, it is basic herein
The upper evaluation index for defining the unbalanced accuracy rate index of classification as feature selecting.Using the unbalanced accuracy rate of classification as target, benefit
Key feature search is carried out with sweep forward method and the learning machine algorithm that transfinites, excavating has stronger identification to unstability class event
The key feature of ability meets the actual demand of electricity net safety stable analysis.
The above is only the preferred embodiment of the present invention, protection scope of the present invention is not limited merely to above-described embodiment,
All technical solutions belonged under thinking of the present invention all belong to the scope of protection of the present invention.It should be pointed out that for the art
For those of ordinary skill, several improvements and modifications without departing from the principles of the present invention should be regarded as protection of the invention
Range.
Claims (6)
1. considering the electric power system transient stability key feature selection method of classification lack of uniformity, which is characterized in that including following
Step:
S1, the primitive character collection that Transient Stability Evaluation is constructed using the time dimension statistic of electric state;
Feature and stable state before and after S2, integration system failure, and generate the sample set of preset quantity;
S3, comprehensive evaluation index class weights accuracy rate CWA is defined according to Transient Stability Evaluation feature selecting interpretational criteria;
S4, firstly, sample set is randomly divided into training set and test set, then, primitive character is concentrated to the classification of single feature
Ability is assessed in training set, and point of the method assessment character subset combined using sweep forward and the learning machine that transfinites
Class performance, and using the maximum character subset of class weights accuracy rate CWA as essential signature sets, finally, utilizing the number of test set
It is tested for the property according to essential signature sets.
2. the electric power system transient stability key feature selection method according to claim 1 for considering classification lack of uniformity,
It is characterized in that, primitive character collection described in step S1 includes the flow state and failure generation before electric power system fault occurs
Electrical response curve afterwards, wherein flow state feature mainly includes the load level of system, generator output level and node
Voltage level;Electrical response feature include failure occur instantaneous and the failure removal moment generator amature angle, rotor velocity,
The condition responsives features such as rotor angular acceleration, active power impact and rotor kinetic energy.
3. the electric power system transient stability key feature selection method according to claim 2 for considering classification lack of uniformity,
It is characterized in that, step S2 further includes following sub-step:
S201, on the basis of benchmark trend, using monte carlo method stochastical sampling generate system load level size, according to
Active power output is adjusted to generator capacity size equal proportion, Load flow calculation is executed and calculation of tidal current is checked, if damp
Stream is not restrained or certain electrical variables are out-of-limit, then the corresponding method of operation is not received, on the contrary then receive the method for operation;
S202, under the feasible method of operation, it is random to set fault disturbance scene, utilize time-domain simulation method to carry out transient stability
Analysis, obtains the transient response track of each generator, and judges the stable state of system after failure generation;
Feature and stable state before and after S203, integration system failure, form transient stability sample;
S204, step S201 to step S203 is repeated, until generating the sample set of preset quantity, and sample set is normalized
Processing.
4. the electric power system transient stability key feature selection method according to claim 3 for considering classification lack of uniformity,
It is characterized in that, Transient Stability Evaluation feature selecting interpretational criteria described in step S3: first is that overall evaluation criterion should be with mistake
Steady sample is unrelated with the relative populations for stablizing sample, avoids the influence of unbalanced data volume;Second is that the weight of unstability sample should be protruded
The property wanted selects the feature for having stronger identification capability to unstability sample.
5. the electric power system transient stability key feature selection method according to claim 4 for considering classification lack of uniformity,
It is characterized in that, comprehensive evaluation index class weights accuracy rate CWA described in step S3 are as follows:
In formula: TSR and TUR is respectively the accuracy rate of Stabilized sample and unstability class sample;α is adjustable factors, steady for characterizing
Determine the significance level of class and unstability class;Wherein, the comprehensive evaluation index class weights accuracy rate index meets transient stability
Assess feature selecting interpretational criteria.
6. the electric power system transient stability key feature selection method according to claim 5 for considering classification lack of uniformity,
It is characterized in that, step S4 further includes following sub-step:
S401, sample set is randomly divided into training set and test set, carries out key feature selection using training set and classifier is instructed
Practice;
S402, it is scanned for using sweep forward method, defined feature subset is empty set first, in training set, using transfiniting
Learning machine concentrates the classification capacity of single feature to assess primitive character, according to the size of class weights accuracy rate value to list
A feature ordering selects class weights accuracy rate to be worth maximum feature and character subset is added;
S403, remaining feature are combined with the feature in character subset one by one, using the learning machine that transfinites to assemblage characteristic classification performance
Assessment selects class weights accuracy rate to be worth maximum spy according to the class weights accuracy rate of assemblage characteristic to remaining feature ordering
Character subset is added in sign;
S404, step S403 is repeated, until character subset has been added in all features, selection corresponds to classification in search process and adds
The power maximum character subset of accuracy rate is essential signature sets;
S405, the essential signature sets obtained according to screening, are tested for the property using the data of test set.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811301951.7A CN109193703A (en) | 2018-11-02 | 2018-11-02 | Consider the electric power system transient stability key feature selection method of classification lack of uniformity |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811301951.7A CN109193703A (en) | 2018-11-02 | 2018-11-02 | Consider the electric power system transient stability key feature selection method of classification lack of uniformity |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109193703A true CN109193703A (en) | 2019-01-11 |
Family
ID=64941715
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811301951.7A Pending CN109193703A (en) | 2018-11-02 | 2018-11-02 | Consider the electric power system transient stability key feature selection method of classification lack of uniformity |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109193703A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110310033A (en) * | 2019-06-28 | 2019-10-08 | 清华大学 | Transient Stability Evaluation method and system based on trend vector Minkowski Distance |
CN112952830A (en) * | 2021-04-16 | 2021-06-11 | 山东科技大学 | Method for predicting transient stability of power system after fault |
CN113516513A (en) * | 2021-07-20 | 2021-10-19 | 重庆度小满优扬科技有限公司 | Data analysis method and device, computer equipment and storage medium |
CN114266396A (en) * | 2021-12-21 | 2022-04-01 | 国网天津市电力公司 | Transient stability discrimination method based on intelligent screening of power grid characteristics |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014179883A1 (en) * | 2013-05-06 | 2014-11-13 | University Of Saskatchewan | Methods and apparatus for detection of transient instability and out-of-step conditions by state deviation |
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 |
CN104881741A (en) * | 2015-05-25 | 2015-09-02 | 清华大学 | Power system transient stability determination method based on support vector machine |
CN106849069A (en) * | 2017-03-13 | 2017-06-13 | 东北电力大学 | A kind of transient stability evaluation in power system method based on Pin SVM |
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
- 2018-11-02 CN CN201811301951.7A patent/CN109193703A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014179883A1 (en) * | 2013-05-06 | 2014-11-13 | University Of Saskatchewan | Methods and apparatus for detection of transient instability and out-of-step conditions by state deviation |
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 |
CN104881741A (en) * | 2015-05-25 | 2015-09-02 | 清华大学 | Power system transient stability determination method based on support vector machine |
CN106849069A (en) * | 2017-03-13 | 2017-06-13 | 东北电力大学 | A kind of transient stability evaluation in power system method based on Pin SVM |
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 (6)
Title |
---|
CHUN ZHANG等: "Feature selection of power system transient stability assessment based on random forest and recursive feature elimination", 《2016 IEEE PES ASIA-PACIFIC POWER AND ENERGY ENGINEERING CONFERENCE 》 * |
于志勇等: "基于GA和ELM的电能质量扰动识别特征选择方法", 《电测与仪表》 * |
刘新东等: "计及不确定性的暂态稳定在线预测", 《中国电机工程学报》 * |
卢锦玲等: "极限学习机和遗传算法在暂态稳定评估特征选择中的应用", 《电力系统及其自动化学报》 * |
李扬等: "基于改进最大相关最小冗余判据的暂态稳定评估特征选择", 《中国电机工程学报》 * |
陈振等: "基于代价敏感极端学习机的电力系统暂态稳定评估方法", 《电力自动化设备》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110310033A (en) * | 2019-06-28 | 2019-10-08 | 清华大学 | Transient Stability Evaluation method and system based on trend vector Minkowski Distance |
CN110310033B (en) * | 2019-06-28 | 2021-10-15 | 清华大学 | Transient stability evaluation method and system based on tidal current vector Minkowski distance |
CN112952830A (en) * | 2021-04-16 | 2021-06-11 | 山东科技大学 | Method for predicting transient stability of power system after fault |
CN113516513A (en) * | 2021-07-20 | 2021-10-19 | 重庆度小满优扬科技有限公司 | Data analysis method and device, computer equipment and storage medium |
CN114266396A (en) * | 2021-12-21 | 2022-04-01 | 国网天津市电力公司 | Transient stability discrimination method based on intelligent screening of power grid characteristics |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109193703A (en) | Consider the electric power system transient stability key feature selection method of classification lack of uniformity | |
Wongwathanarat et al. | Three-dimensional simulations of core-collapse supernovae: from shock revival to shock breakout | |
CN102799822B (en) | Software running security measurement and estimation method based on network environment | |
Donato et al. | Large scale properties of the webgraph | |
Kainulainen et al. | Connection between dense gas mass fraction, turbulence driving, and star formation efficiency of molecular clouds | |
Levitin et al. | Reliability of series-parallel systems with random failure propagation time | |
Han et al. | Probabilistic seismic hazard analysis for spatially distributed infrastructure | |
Hennebelle | The FRIGG project: from intermediate galactic scales to self-gravitating cores | |
Ma et al. | Synthetic population generation with multilevel controls: A fitness‐based synthesis approach and validations | |
Cranmer et al. | Maximum significance at the LHC and Higgs decays to muons | |
CN104536881A (en) | Public testing error report priority sorting method based on natural language analysis | |
CN106682152A (en) | Recommendation method for personalized information | |
Reiss et al. | Machine learning for predicting the Bz magnetic field component from upstream in situ observations of solar coronal mass ejections | |
Eldridge et al. | Supernova lightCURVE POPulation Synthesis II: Validation against supernovae with an observed progenitor | |
Tjus et al. | Cosmic-ray propagation around the Sun: investigating the influence of the solar magnetic field on the cosmic-ray Sun shadow | |
Pandya et al. | Variation of radiation belt electron flux during CME‐and CIR‐driven geomagnetic storms: Van Allen Probes observations | |
De Francesco et al. | A proposal to update LSA databases for an operational availability based on autonomic logistic | |
Etesami et al. | Searching for lepton flavor violating interactions at future electron-positron colliders | |
Anderson et al. | Dynamical system decomposition using dissipation inequalities | |
CN110428164A (en) | A kind of operation of power networks state monitoring method, device, equipment and readable storage medium storing program for executing | |
Shi et al. | Skyline‐based ground motion selection method for nonlinear time history analysis of building structures | |
Wang | Search for Magnetic Monopoles with the NO $\nu $ A Far Detector | |
Chekanov et al. | Bin-bin correlation measurement by the bunching-parameter method | |
Riaz et al. | An Artificial Neural Network Based Digital Differential Protection Scheme for Synchronous Generator Stator Winding Protection | |
Di Florio | Universita degli Studi di Bari |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190111 |