CN110108992A - Based on cable partial discharge fault recognition method, system and the medium for improving random forests algorithm - Google Patents
Based on cable partial discharge fault recognition method, system and the medium for improving random forests algorithm Download PDFInfo
- Publication number
- CN110108992A CN110108992A CN201910440682.0A CN201910440682A CN110108992A CN 110108992 A CN110108992 A CN 110108992A CN 201910440682 A CN201910440682 A CN 201910440682A CN 110108992 A CN110108992 A CN 110108992A
- Authority
- CN
- China
- Prior art keywords
- cable
- partial discharge
- random forests
- forests algorithm
- attribute
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
- G01R31/1263—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation
- G01R31/1272—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation of cable, line or wire insulation, e.g. using partial discharge measurements
Abstract
The invention discloses a kind of based on the cable partial discharge fault recognition method, system and the medium that improve random forests algorithm, and the method for the present invention includes the local discharge signal for acquiring cable；Extract the characteristic of local discharge signal；Obtained characteristic input will be extracted to construct in advance and complete trained classifier, obtain the corresponding cable partial discharge fault type of cable local discharge signal, classifier is based on the classifier for improving random forests algorithm, the selection that random forests algorithm excavates higher-dimension attribute data and the data weighting amendment thought guidance attribute set based on Adaboost algorithm with feature construction method is improved, and classifier is established the mapping relations between the feature of local discharge signal, cable partial discharge fault type by preparatory training.The present invention can guarantee to also improve the efficiency of identification while identifying accuracy, realize that identification accuracy and efficiency has both.
Description
Technical field
The present invention relates to cable ageing management fields, and in particular to a kind of based on the cable partial discharge for improving random forests algorithm
Fault recognition method, system and medium.
Background technique
Electric energy play the role of in real life it is indispensable, and cable as transmission electric energy important component,
It is widely used, main status is especially occupied in urban power distribution network.The part of cable generation shelf depreciation
Discharge capacity and its current insulation status have close relationship, therefore evaluate the most effective, most intuitive of cable insulation situation
Mode is exactly the partial discharge quantity for measuring cable.Therefore, in perfect operation cable local discharge diagnostic method, to improve electric power
The reliability of system and safety are of great significance.
Cable can be with physical phenomenons such as light, heat, electromagnetic wave, electric pulses when generating shelf depreciation, these physical phenomenons are just
It is the foundation of cable local discharge detection.Although from cable generate shelf depreciation in it will be seen that cable insulation status,
It is the defect type of cable local discharge failure to be differentiated, needs to carry out pattern-recognition to local discharge signal.Part is put at present
The main method of power mode identification is shelf depreciation type several frequently seen according to cable first, and the test of corresponding types is arranged,
And largely tested, characteristic is then extracted from the test of every set type again, and establish to obtained characteristic
Spectrum library is finally trained obtained a few class spectrum datas using intelligent algorithm, is finally reached the effect of Classification and Identification.?
The intelligent algorithm being commonly used during Classification and Identification includes mode random forest grader, supporting vector based on distance
Machine, neural network and fuzzy stochastic forest classified device etc..Under the premise of a large amount of training data and training time, these intelligence
Energy algorithm can obtain good effect, but PD Pattern Recognition tends not in accuracy rate and on the time get both.
Summary of the invention
The technical problem to be solved in the present invention: in view of the above problems in the prior art, providing one kind can guarantee to identify
Also improved while accuracy identification efficiency, realize identification accuracy and efficiency have both based on improve random forests algorithm
Cable partial discharge fault recognition method, system and medium.
In order to solve the above-mentioned technical problem, the technical solution adopted by the present invention are as follows:
A kind of cable partial discharge fault recognition method based on improvement random forests algorithm, implementation steps include:
1) local discharge signal of cable is acquired；
2) characteristic of local discharge signal is extracted；
3) obtained characteristic input will be extracted and constructs and complete trained classifier in advance, obtain the cable part
The corresponding cable partial discharge fault type of discharge signal, the classifier is the classifier based on improvement random forests algorithm, described
Random forests algorithm is improved to repair with feature construction method excavation higher-dimension attribute data and the data weighting based on Adaboost algorithm
The selection of positive thought guidance attribute set, and the classifier is established feature, the cable of local discharge signal by preparatory training
Mapping relations between partial discharge fault type.
Preferably, the characteristic in step 2) includes at least one of 24 high dimension attributes, and 24 high dimension attributes include WithThe degree of asymmetry Asy and cross-correlation coefficient Cc of three, WithDegree of skewness Sk, steepness Ku, peak value Peak；Its
In,Indicate mean discharge magnitude phase distribution two dimension spectrogram,Indicate maximum pd quantity phase distribution two-dimensional spectrum
Figure,Indicate discharge time phase distribution spectrogram；Indicate positive half cycle mean discharge magnitude phase distribution two dimension spectrogram,Indicate negative half period mean discharge magnitude phase distribution two dimension spectrogram,Indicate positive half cycle maximum pd quantity phase point
Cloth two dimension spectrogram,Indicate negative half period maximum pd quantity phase distribution two dimension spectrogram,Indicate positive half cycle electric discharge time
Number phase distribution spectrogram,Indicate negative half period discharge time phase distribution spectrogram.
Preferably, the cable partial discharge fault type includes internal discharge, creeping discharge, corona discharge.
Preferably, further include the steps that trained classifier, detailed step include: before step 3)
S1 cable local discharge signal) is acquired for various cable partial discharge fault types respectively and extracts characteristic, point
A part of it Xuan Qu not be used as training set, another part is as test set；Training set is adopted again using Bootstrap method
Training set is randomly generated in sample, and wherein the attribute sum of training set is M, and attribute sum is corresponding to difference judging characteristic data
Judging basis sum, the number of sub- attribute are N；
S2) using the selection of the data set weight amendment thought guidance attribute set of Adaboost algorithm, to attribute set
Optimize processing, the attribute set is the set of judging basis corresponding to characteristic parameter used in assorting process；?
The selection of attribute set Q is carried out on the basis of above-mentioned training set, training set is characteristic parameter used in assorting process, sub- category
Property is the corresponding judging basis of characteristic parameter；And category is directly modified by the fault data in the Ci of a upper sub-tree
Temper collection Q (i+1) selected probability, sub-tree are the visualization descriptions to every group of data from classification start and ending, according to
Probability value after finishing completes the selection of attribute set according to conventional random device；
S3 corresponding decision tree) is generated using the training set and corresponding attribute set that construct, is to draw with information gain
Divide principle, the divisional mode when information gain obtains maximum is exactly the best divisional mode of such decision tree, and in this way
Corresponding node is divided；And each tree is all completely grown up, and without beta pruning；
S4) test set is tested using decision tree, obtains corresponding classification, most classifications will be exported in decision tree
As classification belonging to the test set, to establish reflecting between the feature of local discharge signal, cable partial discharge fault type
Penetrate relationship.
It is set the present invention also provides a kind of based on the electric cable stoppage identifying system for improving random forests algorithm, including computer
Standby, which is programmed or configures aforementioned based on the cable partial discharge failure for improving random forests algorithm to execute the present invention
The step of recognition methods.
It is set the present invention also provides a kind of based on the electric cable stoppage identifying system for improving random forests algorithm, including computer
It is standby, be stored on the storage medium of the computer equipment be programmed or configure with execute the present invention it is aforementioned based on improve random forest
The computer program of the cable partial discharge fault recognition method of algorithm.
The present invention also provides a kind of based on the electric cable stoppage identifying system for improving random forests algorithm, including interconnected
Partial discharge signal acquires equipment and host computer, which is programmed or configures aforementioned based on improvement random forest to execute the present invention
The step of cable partial discharge fault recognition method of algorithm.
The present invention also provides a kind of based on the electric cable stoppage identifying system for improving random forests algorithm, including interconnected
Partial discharge signal acquires equipment and host computer, is stored with before being programmed or configuring to execute the present invention on the storage medium of the host computer
State the computer program based on the cable partial discharge fault recognition method for improving random forests algorithm.
The present invention also provides a kind of computer readable storage medium, it is stored with and is programmed on the computer readable storage medium
Or configuration is to execute the aforementioned computer program based on the cable partial discharge fault recognition method for improving random forests algorithm of the present invention.
The present invention also provides a kind of based on the electric cable stoppage identifying system for improving random forests algorithm, comprising:
Signal acquisition program unit, for acquiring the local discharge signal of cable；
Feature extraction program unit, for extracting the characteristic of local discharge signal；
Classification and Identification program unit, for the preparatory classification constructed and complete training of obtained characteristic input will to be extracted
Device, obtains the corresponding cable partial discharge fault type of the cable local discharge signal, and the classifier is random gloomy based on improving
The classifier of woods algorithm, the improvement random forests algorithm excavate higher-dimension attribute data with feature construction method and are based on
The selection of the data weighting amendment thought guidance attribute set of Adaboost algorithm, and the classifier is established by preparatory training
Mapping relations between the feature of local discharge signal, cable partial discharge fault type.
With the prior art, the present invention has an advantage that the present invention inputs preparatory structure for obtained characteristic is extracted
Build and complete trained classifier, obtain the corresponding cable partial discharge fault type of cable local discharge signal, classifier be based on
The classifier of random forests algorithm is improved, random forests algorithm is improved and excavates higher-dimension attribute data with feature construction method and be based on
The selection of the data weighting amendment thought guidance attribute set of Adaboost algorithm, and classifier establishes part by preparatory training
Mapping relations between the feature of discharge signal, cable partial discharge fault type, classifier is using improving random forests algorithm to electricity
Cable partial discharges fault type is identified, excavates higher-dimension attribute data by utilization feature construction method, and be based on
The data set weight of Adaboost algorithm corrects thought, the selection of attribute set is instructed, to cable local discharge fault type
Identification guarantees to substantially reduce the time in the case where accuracy, improves recognition efficiency.
Detailed description of the invention
Fig. 1 is the basic procedure schematic diagram of present invention method.
Fig. 2 is the cable local discharge fault type recognition of 300 groups of test datas in the embodiment of the present invention by distribution map.
Specific embodiment
As shown in Figure 1, implementation step of the present embodiment based on the cable partial discharge fault recognition method for improving random forests algorithm
Suddenly include:
1) local discharge signal of cable is acquired；
2) characteristic of local discharge signal is extracted；
3) obtained characteristic input will be extracted and constructs and complete trained classifier in advance, obtain cable local discharge
The corresponding cable partial discharge fault type of signal, classifier are to improve random forest based on the classifier for improving random forests algorithm
Algorithm corrects thought guidance category with feature construction method excavation higher-dimension attribute data and the data weighting based on Adaboost algorithm
The selection of temper collection, and classifier is established between the feature of local discharge signal, cable partial discharge fault type by preparatory training
Mapping relations.
In the present embodiment, the characteristic in step 2) includes 24 high dimension attributes, referring to table 1:
As shown in table 1, above-mentioned 24 high dimension attributes includeWithThe degree of asymmetry Asy of three
And cross-correlation coefficient Cc,WithDegree of skewness Sk, steep
Kurtosis Ku, peak value Peak；Wherein,Indicate mean discharge magnitude phase distribution two dimension spectrogram,Indicate maximum electric discharge
Phase distribution two dimension spectrogram is measured,Indicate discharge time phase distribution spectrogram；Indicate positive half cycle mean discharge magnitude
Phase distribution two dimension spectrogram,Indicate negative half period mean discharge magnitude phase distribution two dimension spectrogram,Indicate just half
All maximum pd quantity phase distribution two dimension spectrograms,Indicate negative half period maximum pd quantity phase distribution two dimension spectrogram,Indicate positive half cycle discharge time phase distribution spectrogram,Indicate negative half period discharge time phase distribution spectrogram.This
Outside, also can according to need one of aforementioned 24 attributes of selection or it is a variety of come construction feature data, only from accuracy
Speech, the attribute of use is more, then the characteristic completeness constructed is better, and cable partial discharge fault identification result is more accurate.Step
3) it specifically refers to construct phase using basic Partial Discharge as object with feature construction method excavation higher-dimension attribute data in
The characteristic answered, i.e. 24 high dimension attributes shown in table 1.
In the present embodiment, cable partial discharge fault type includes internal discharge, creeping discharge, corona discharge.
In the present embodiment, further include the steps that trained classifier, detailed step include: before step 3)
S1 cable local discharge signal) is acquired for various cable partial discharge fault types respectively and extracts characteristic, point
A part of it Xuan Qu not be used as training set, another part is as test set；Training set is adopted again using Bootstrap method
Training set is randomly generated in sample, and wherein the attribute sum of training set is M, and attribute sum is corresponding to difference judging characteristic data
Judging basis sum, the number of sub- attribute are N；
S2) using the selection of the data set weight amendment thought guidance attribute set of Adaboost algorithm, to attribute set
Optimize processing, attribute set is the set of judging basis corresponding to characteristic parameter used in assorting process；Above-mentioned
Training set on the basis of carry out attribute set Q selection, training set is characteristic parameter used in assorting process, and sub- attribute is
The corresponding judging basis of characteristic parameter；And attribute is directly modified by the fault data in the Ci of a upper sub-tree
Collect Q (i+1) selected probability, sub-tree is the visualization description to every group of data from classification start and ending, according to finishing
Probability value afterwards completes the selection of attribute set according to conventional random device；
In the present embodiment, by cable internal discharge, creeping discharge, corona discharge three types partial discharge test,
After a large number of experiments, 600 groups of characteristics with 24 attributes are extracted to every group of test respectively, obtain 1800 groups of characteristics
Afterwards, 1500 groups of characteristics are randomly selected as training data, are left 300 groups of data as test set.Create random forest point
Class device: with Bootstrap method to training data resampling, 1500 training sets, S are obtained1, S2, S3..., S1500, belong to
Property sum be M, the number of sub- attribute is N.
The selection of attribute set Q, the C of a upper sub-tree are carried out on the basis of 1500 above-mentioned training setsiIn
Fault data directly affect the selected probability of an attribute set Q (i+1), each attribute is q in attribute seti,jBelong to for j-th
Property qjSelected probability value, wherein j=1,2,3 ..., m；.Assuming that CiIn sub-tree in training subset diagnostic error data set
It is combined into mi, as training subset Si+1In belong to set miIn data amount check be niWhen, then next attribute set Qi+1In each attribute
Selected probability P change.In the present embodiment, attribute is directly modified by the fault data in the Ci of a upper sub-tree
The calculating function expression of subset Q (i+1) selected probability P are as follows:
In above formula, M is the attribute sum of training set, and N is the number of sub- attribute, and k is middle space variable, and calculation is such as
Under:
In above formula, niFor next training subset Si+1In belong to a troubleshooting set miIn data amount check, mi
For CiIn sub-tree in training subset diagnostic error data acquisition system；qi,jFor j-th of attribute qjSelected probability value, wherein j
=1,2,3 ..., m；PjFor j-th of attribute qjSelected probability value；M is attribute sum；N is attribute number in sub- property set；Qi
For the property set of a upper sub-tree；α is the accuracy rate threshold value of decision tree training.
S3) using the training set that constructs and corresponding attribute set generate corresponding decision tree (C1, C2, C3 ...,
C1500), using information gain as division principle, when information gain obtains it is maximum when divisional mode be exactly such decision tree most
Good divisional mode, and corresponding node is divided in this way；And each tree is all completely grown up, and without beta pruning；
In the present embodiment, corresponding decision tree C is generated using each training set1、C2、C3、…、C1500, in each non-leaf
Before selecting attribute on node, Split Attribute collection of the m attribute as present node is randomly selected from 24 attributes, m < 24, and
The node is divided with divisional mode best in this m attribute, each tree is all completely grown up, and without beta pruning.
Using information gain as division principle, the specific steps of which are as follows:
M data is divided into n class, the ratio of every one kind is PiNumber/m of=the i-th class；
If D is sample training collection, then the entropy inf (D) of sample training collection D are as follows:
In above formula, PiFor the ratio that m data is divided into the i-th class that n class obtains, n is that data divide class number.
Assuming that dividing the sample in sample training collection D according to attribute A, sample training collection D is divided into v difference by attribute A
Class, the entropy inf of sample training collection D after divisionA(D) are as follows:
In above formula, DjFor the number of jth class in the V class of division, v is that the classification that attribute A divides sample training collection D is combed
Reason, Info (Dj) it is DjEntropy, D be sample training collection D.
It follows that information gain Gain (A) are as follows:
Gain (A)=inf (D)-infA(D)
In above formula, inf (D) is the entropy of sample training collection D, infA(D) for according in attribute A division sample training collection D
The entropy of this training set D after sample, the two are subtracted each other as information gain Gain (A).
S4) test set is tested using decision tree, obtains corresponding classification, most classifications will be exported in decision tree
As classification belonging to the test set, to establish reflecting between the feature of local discharge signal, cable partial discharge fault type
Penetrate relationship.In the present embodiment, using the data to be tested such as individual as test set sample X, surveyed using each decision tree
Examination is tested using 1500 decision trees that random forest method is established, obtains corresponding classification C1 (X), C2 (X), C3
(X) ..., C1500 (X) will export most classifications as test set sample X using the method for ballot in 1500 decision trees
Affiliated classification, to establish the mapping relations between the feature of local discharge signal, cable partial discharge fault type.
Fig. 2 is the cable local discharge fault type recognition of 300 groups of test datas in the embodiment of the present invention by distribution map, is joined
See Fig. 2 it is found that in the recognition result of 300 groups of test datas, internal discharge (bubble-discharge), creeping discharge, corona discharge three
In kind decision tree classification, internal discharge (bubble-discharge) decision tree number is most, thus identify that final cable partial discharge failure
Type is internal discharge (bubble-discharge).Using the classifier in the present embodiment to partial discharge of transformer Fault Pattern Recognition,
The time can be substantially reduced while guaranteeing accuracy rate.
In addition, the present embodiment also provides a kind of electric cable stoppage identifying system based on improvement random forests algorithm, including meter
Machine equipment is calculated, which is programmed or configures aforementioned based on the cable for improving random forests algorithm to execute the present embodiment
The step of partial discharge fault recognition method.In addition, the present embodiment also provides a kind of electric cable stoppage based on improvement random forests algorithm
Identifying system, including computer equipment are stored on the storage medium of the computer equipment and are programmed or configure to execute this reality
Apply the aforementioned computer program based on the cable partial discharge fault recognition method for improving random forests algorithm of example.In addition, the present embodiment
A kind of electric cable stoppage identifying system based on improvement random forests algorithm is also provided, including Partial discharge signal interconnected acquisition is set
Standby and host computer, the host computer are programmed or configure aforementioned based on the cable office for improving random forests algorithm to execute the present embodiment
The step of putting fault recognition method.In addition, the present embodiment also provides a kind of electric cable stoppage knowledge based on improvement random forests algorithm
Other system, including Partial discharge signal interconnected acquire equipment and host computer, are stored with and are compiled on the storage medium of the host computer
Journey or configuration are to execute the aforementioned computer based on the cable partial discharge fault recognition method for improving random forests algorithm of the present embodiment
Program.In addition, the present embodiment also provides a kind of computer readable storage medium, be stored on the computer readable storage medium by
Programming is configured to execute the aforementioned calculating based on the cable partial discharge fault recognition method for improving random forests algorithm of the present embodiment
Machine program.In addition, the present embodiment also provides a kind of electric cable stoppage identifying system based on improvement random forests algorithm, comprising:
Signal acquisition program unit, for acquiring the local discharge signal of cable；
Feature extraction program unit, for extracting the characteristic of local discharge signal；
Classification and Identification program unit, for the preparatory classification constructed and complete training of obtained characteristic input will to be extracted
Device, obtains the corresponding cable partial discharge fault type of the cable local discharge signal, and the classifier is random gloomy based on improving
The classifier of woods algorithm, the improvement random forests algorithm excavate higher-dimension attribute data with feature construction method and are based on
The selection of the data weighting amendment thought guidance attribute set of Adaboost algorithm, and the classifier is established by preparatory training
Mapping relations between the feature of local discharge signal, cable partial discharge fault type.
It should be understood by those skilled in the art that, embodiments herein can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the application, which can be used in one or more,
The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces
The form of product.
The application is referring to method, the process of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions
The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs
Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real
The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.
The above is only a preferred embodiment of the present invention, protection scope of the present invention is not limited merely to above-mentioned implementation
Example, 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
Those of ordinary skill for, several improvements and modifications without departing from the principles of the present invention, these improvements and modifications
It should be regarded as protection scope of the present invention.
Claims (10)
1. a kind of based on the cable partial discharge fault recognition method for improving random forests algorithm, it is characterised in that implementation steps include:
1) local discharge signal of cable is acquired；
2) characteristic of local discharge signal is extracted；
3) obtained characteristic input will be extracted and constructs and complete trained classifier in advance, obtain the cable local discharge
The corresponding cable partial discharge fault type of signal, the classifier are based on the classifier for improving random forests algorithm, the improvement
Random forests algorithm is corrected and is thought with feature construction method excavation higher-dimension attribute data and the data weighting based on Adaboost algorithm
It is intended to refer to the selection for leading attribute set, and the classifier is established feature, the cable partial discharge of local discharge signal by preparatory training
Mapping relations between fault type.
2. according to claim 1 based on the cable partial discharge fault recognition method for improving random forests algorithm, feature exists
In the characteristic in step 2) includes at least one of 24 high dimension attributes, and 24 high dimension attributes includeWithThe degree of asymmetry Asy and cross-correlation coefficient Cc of three,WithDegree of skewness Sk, steepness Ku, peak value Peak；Its
In,Indicate mean discharge magnitude phase distribution two dimension spectrogram,Indicate maximum pd quantity phase distribution two-dimensional spectrum
Figure,Indicate discharge time phase distribution spectrogram；Indicate positive half cycle mean discharge magnitude phase distribution two dimension spectrogram,Indicate negative half period mean discharge magnitude phase distribution two dimension spectrogram,Indicate positive half cycle maximum pd quantity phase point
Cloth two dimension spectrogram,Indicate negative half period maximum pd quantity phase distribution two dimension spectrogram,Indicate positive half cycle electric discharge time
Number phase distribution spectrogram,Indicate negative half period discharge time phase distribution spectrogram.
3. according to claim 1 based on the cable partial discharge fault recognition method for improving random forests algorithm, feature exists
In the cable partial discharge fault type includes internal discharge, creeping discharge, corona discharge.
4. according to claim 1 based on the cable partial discharge fault recognition method for improving random forests algorithm, feature exists
In step 3) further includes the steps that trained classifier, detailed step include: before
S1 cable local discharge signal) is acquired for various cable partial discharge fault types respectively and extracts characteristic, is selected respectively
Take a part as training set, another part is as test set；Resampling is carried out using Bootstrap method to training set, with
Machine generate training set, wherein the attribute sum of training set be M, attribute sum be distinguish judging characteristic data corresponding to judge according to
According to sum, the number of sub- attribute is N；
S2) using the selection of the data set weight amendment thought guidance attribute set of Adaboost algorithm, attribute set is carried out
Optimization processing, the attribute set are the set of judging basis corresponding to characteristic parameter used in assorting process；Above-mentioned
Training set on the basis of carry out attribute set Q selection, training set is characteristic parameter used in assorting process, and sub- attribute is
The corresponding judging basis of characteristic parameter；And attribute is directly modified by the fault data in the Ci of a upper sub-tree
Collect Q (i+1) selected probability, sub-tree is the visualization description to every group of data from classification start and ending, according to finishing
Probability value afterwards completes the selection of attribute set according to conventional random device；
S3 corresponding decision tree) is generated using the training set and corresponding attribute set that construct, is to divide original with information gain
Then, the divisional mode when information gain obtains maximum is exactly the best divisional mode of such decision tree, and in this way to right
The node answered is divided；And each tree is all completely grown up, and without beta pruning；
S4) test set is tested using decision tree, obtains corresponding classification, will be exported in decision tree most classifications as
Classification belonging to the test set, so that the mapping established between the feature of local discharge signal, cable partial discharge fault type is closed
System.
5. a kind of based on the electric cable stoppage identifying system for improving random forests algorithm, including computer equipment, which is characterized in that should
Computer equipment be programmed or configure with perform claim require any one of 1~4 described in based on improving random forests algorithm
The step of cable partial discharge fault recognition method.
6. a kind of based on the electric cable stoppage identifying system for improving random forests algorithm, including computer equipment, which is characterized in that should
It is stored on the storage medium of computer equipment and is programmed or configures to be based on described in any one of perform claim requirement 1~4
Improve the computer program of the cable partial discharge fault recognition method of random forests algorithm.
7. a kind of set based on the electric cable stoppage identifying system for improving random forests algorithm, including Partial discharge signal interconnected acquisition
Standby and host computer, which is characterized in that the host computer is programmed or is configured with base described in any one of perform claim requirement 1~4
In the step of improving the cable partial discharge fault recognition method of random forests algorithm.
8. a kind of set based on the electric cable stoppage identifying system for improving random forests algorithm, including Partial discharge signal interconnected acquisition
Standby and host computer, which is characterized in that be stored on the storage medium of the host computer be programmed or configure with perform claim require 1~
Computer program based on the cable partial discharge fault recognition method for improving random forests algorithm described in any one of 4.
9. a kind of computer readable storage medium, which is characterized in that be stored on the computer readable storage medium be programmed or
Configuration is with the cable partial discharge fault identification side based on improvement random forests algorithm described in any one of perform claim requirement 1~4
The computer program of method.
10. a kind of based on the electric cable stoppage identifying system for improving random forests algorithm characterized by comprising
Signal acquisition program unit, for acquiring the local discharge signal of cable；
Feature extraction program unit, for extracting the characteristic of local discharge signal；
Classification and Identification program unit constructs and completes trained classifier in advance for will extract the input of obtained characteristic,
The corresponding cable partial discharge fault type of the cable local discharge signal is obtained, the classifier is to calculate based on improvement random forest
The classifier of method, the improvement random forests algorithm are excavated higher-dimension attribute data with feature construction method and are calculated based on Adaboost
The selection of the data weighting amendment thought guidance attribute set of method, and the classifier establishes shelf depreciation by preparatory training and believes
Number feature, the mapping relations between cable partial discharge fault type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910440682.0A CN110108992B (en) | 2019-05-24 | 2019-05-24 | Cable partial discharge fault identification method and system based on improved random forest algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910440682.0A CN110108992B (en) | 2019-05-24 | 2019-05-24 | Cable partial discharge fault identification method and system based on improved random forest algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110108992A true CN110108992A (en) | 2019-08-09 |
| CN110108992B CN110108992B (en) | 2021-07-23 |
Family
ID=67492128
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910440682.0A Active CN110108992B (en) | 2019-05-24 | 2019-05-24 | Cable partial discharge fault identification method and system based on improved random forest algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110108992B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110796187A (en) * | 2019-10-22 | 2020-02-14 | 西安奕斯伟硅片技术有限公司 | Method and device for classifying defects |
| CN111366816A (en) * | 2020-04-26 | 2020-07-03 | 华北电力大学 | Power grid fault diagnosis method based on machine learning |
| CN111721834A (en) * | 2020-06-22 | 2020-09-29 | 南京南瑞继保电气有限公司 | Cable partial discharge online monitoring defect identification method |
| CN111815027A (en) * | 2020-06-09 | 2020-10-23 | 山东大学 | Photovoltaic station generated power prediction method and system |
| CN112308146A (en) * | 2020-11-02 | 2021-02-02 | 国网福建省电力有限公司 | Distribution transformer fault identification method based on operation characteristics |
| CN113128568A (en) * | 2021-03-26 | 2021-07-16 | 广州大学 | Excavator activity identification method, system, device and storage medium |
| CN113283371A (en) * | 2021-06-09 | 2021-08-20 | 上海交通大学 | Local discharge feature extraction and classification method based on BRISK features |
| CN114167238A (en) * | 2021-12-03 | 2022-03-11 | 广东电网有限责任公司广州供电局 | Cable partial discharge detection method, device, system, equipment and storage medium |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923650A (en) * | 2010-08-27 | 2010-12-22 | 北京大学 | Random forest classification method and classifiers based on comparison mode |
| CN101964063A (en) * | 2010-09-14 | 2011-02-02 | 南京信息工程大学 | Method for constructing improved AdaBoost classifier |
| CN102928720A (en) * | 2012-11-07 | 2013-02-13 | 广东电网公司 | Defect rate detecting method of oil immersed type main transformer |
| CN102937694A (en) * | 2012-11-27 | 2013-02-20 | 广州供电局有限公司 | Device for monitoring external insulation strength of dirty insulator |
| CN103226835A (en) * | 2013-04-15 | 2013-07-31 | 北京大学深圳研究生院 | Target tracking method and system based on on-line initialization gradient enhancement regression tree |
| US20140044342A1 (en) * | 2012-08-07 | 2014-02-13 | Research & Business Foundation Sungkyunkwan University | Method for generating 3d coordinates and mobile terminal for generating 3d coordinates |
| CN104391970A (en) * | 2014-12-04 | 2015-03-04 | 深圳先进技术研究院 | Attribute subspace weighted random forest data processing method |
| CN105759177A (en) * | 2016-04-26 | 2016-07-13 | 浙江大学城市学院 | Classified-multi-mode-fusion-based distributed grid island detection method |
| CN105844300A (en) * | 2016-03-24 | 2016-08-10 | 河南师范大学 | Optimized classification method and optimized classification device based on random forest algorithm |
| CN106778853A (en) * | 2016-12-07 | 2017-05-31 | 中南大学 | Unbalanced data sorting technique based on weight cluster and sub- sampling |
| CN107516108A (en) * | 2017-08-15 | 2017-12-26 | 国网四川省电力公司电力科学研究院 | Grader creation method and partial discharge of transformer method of fault pattern recognition |
| US20190156159A1 (en) * | 2017-11-20 | 2019-05-23 | Kavya Venkata Kota Sai KOPPARAPU | System and method for automatic assessment of cancer |
-
2019
- 2019-05-24 CN CN201910440682.0A patent/CN110108992B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923650A (en) * | 2010-08-27 | 2010-12-22 | 北京大学 | Random forest classification method and classifiers based on comparison mode |
| CN101964063A (en) * | 2010-09-14 | 2011-02-02 | 南京信息工程大学 | Method for constructing improved AdaBoost classifier |
| US20140044342A1 (en) * | 2012-08-07 | 2014-02-13 | Research & Business Foundation Sungkyunkwan University | Method for generating 3d coordinates and mobile terminal for generating 3d coordinates |
| CN102928720A (en) * | 2012-11-07 | 2013-02-13 | 广东电网公司 | Defect rate detecting method of oil immersed type main transformer |
| CN102937694A (en) * | 2012-11-27 | 2013-02-20 | 广州供电局有限公司 | Device for monitoring external insulation strength of dirty insulator |
| CN103226835A (en) * | 2013-04-15 | 2013-07-31 | 北京大学深圳研究生院 | Target tracking method and system based on on-line initialization gradient enhancement regression tree |
| CN104391970A (en) * | 2014-12-04 | 2015-03-04 | 深圳先进技术研究院 | Attribute subspace weighted random forest data processing method |
| CN105844300A (en) * | 2016-03-24 | 2016-08-10 | 河南师范大学 | Optimized classification method and optimized classification device based on random forest algorithm |
| CN105759177A (en) * | 2016-04-26 | 2016-07-13 | 浙江大学城市学院 | Classified-multi-mode-fusion-based distributed grid island detection method |
| CN106778853A (en) * | 2016-12-07 | 2017-05-31 | 中南大学 | Unbalanced data sorting technique based on weight cluster and sub- sampling |
| CN107516108A (en) * | 2017-08-15 | 2017-12-26 | 国网四川省电力公司电力科学研究院 | Grader creation method and partial discharge of transformer method of fault pattern recognition |
| US20190156159A1 (en) * | 2017-11-20 | 2019-05-23 | Kavya Venkata Kota Sai KOPPARAPU | System and method for automatic assessment of cancer |
Non-Patent Citations (2)
| Title |
|---|
| 郭一帆 等: "基于机器学习算法的寿命预测与故障诊断技术的发展综述", 《计算机测量与控制》 * |
| 龙震泽: "不同气隙尺寸的油纸绝缘气隙放电特征及发展阶段识别", 《中国优秀硕士学位论文全文数据库工程科技二辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110796187A (en) * | 2019-10-22 | 2020-02-14 | 西安奕斯伟硅片技术有限公司 | Method and device for classifying defects |
| CN111366816A (en) * | 2020-04-26 | 2020-07-03 | 华北电力大学 | Power grid fault diagnosis method based on machine learning |
| CN111815027A (en) * | 2020-06-09 | 2020-10-23 | 山东大学 | Photovoltaic station generated power prediction method and system |
| CN111721834A (en) * | 2020-06-22 | 2020-09-29 | 南京南瑞继保电气有限公司 | Cable partial discharge online monitoring defect identification method |
| CN111721834B (en) * | 2020-06-22 | 2023-08-08 | 南京南瑞继保电气有限公司 | Cable partial discharge on-line monitoring defect identification method |
| CN112308146A (en) * | 2020-11-02 | 2021-02-02 | 国网福建省电力有限公司 | Distribution transformer fault identification method based on operation characteristics |
| CN113128568A (en) * | 2021-03-26 | 2021-07-16 | 广州大学 | Excavator activity identification method, system, device and storage medium |
| CN113128568B (en) * | 2021-03-26 | 2023-08-11 | 广州大学 | Excavator activity recognition method, system, device and storage medium |
| CN113283371A (en) * | 2021-06-09 | 2021-08-20 | 上海交通大学 | Local discharge feature extraction and classification method based on BRISK features |
| CN114167238A (en) * | 2021-12-03 | 2022-03-11 | 广东电网有限责任公司广州供电局 | Cable partial discharge detection method, device, system, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110108992B (en) | 2021-07-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110108992A (en) | Based on cable partial discharge fault recognition method, system and the medium for improving random forests algorithm | |
| CN107273920A (en) | A kind of non-intrusion type household electrical appliance recognition methods based on random forest | |
| CN110488161A (en) | A kind of detection of multi-load series arc faults and localization method | |
| CN109507535B (en) | Method and device for predicting operation stage and operation life of transformer substation grounding grid | |
| CN107516108A (en) | Grader creation method and partial discharge of transformer method of fault pattern recognition | |
| CN109145957A (en) | The identification and processing method and processing device of power distribution network abnormal index based on big data | |
| CN109765447A (en) | A kind of intelligent substation relay protection automatic test approach | |
| CN106452825A (en) | Power distribution and utilization communication network alarm correlation analysis method based on improved decision tree | |
| CN105388402A (en) | Cross-linked cable partial discharge mode recognition method based on parameter optimization SVM (Support Vector Machine) algorithm | |
| CN103675610A (en) | Method for extracting characteristic factors in online local discharge detection | |
| CN107729939A (en) | A kind of CIM extended method and device towards newly-increased power network resources | |
| CN108989581A (en) | A kind of consumer's risk recognition methods, apparatus and system | |
| CN107784392A (en) | A kind of the defects of transmission line of electricity based on machine learning Forecasting Methodology and device | |
| CN108805295A (en) | A kind of method for diagnosing faults based on decision Tree algorithms | |
| CN105930900A (en) | Method and system for predicting hybrid wind power generation | |
| CN115618249A (en) | Low-voltage power distribution station area phase identification method based on LargeVis dimension reduction and DBSCAN clustering | |
| Dasgupta et al. | Broad toplike vector quarks at LHC and HL-LHC | |
| CN108446712A (en) | ODN nets Intelligent planning method, apparatus and system | |
| CN110378358A (en) | A kind of power distribution network isomeric data integration method and system | |
| CN107423328A (en) | The construction method of big data electric power first-aid hotspot prediction system | |
| CN105701265B (en) | A kind of double-fed fan motor unit modeling method and device | |
| CN107276093B (en) | The Probabilistic Load calculation method cut down based on scene | |
| CN116401532B (en) | Method and system for recognizing frequency instability of power system after disturbance | |
| Gerster et al. | Comparison of random sampling and heuristic optimization-based methods for determining the flexibility potential at vertical system interconnections | |
| CN109670526A (en) | A kind of interference source type discrimination method, device, equipment and the storage medium of power grid |
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 |