CN112485610A - GIS partial discharge characteristic parameter extraction method considering insulation degradation - Google Patents
GIS partial discharge characteristic parameter extraction method considering insulation degradation Download PDFInfo
- Publication number
- CN112485610A CN112485610A CN202011226040.XA CN202011226040A CN112485610A CN 112485610 A CN112485610 A CN 112485610A CN 202011226040 A CN202011226040 A CN 202011226040A CN 112485610 A CN112485610 A CN 112485610A
- Authority
- CN
- China
- Prior art keywords
- feature
- insulation
- gis
- partial discharge
- characteristic parameter
- 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
- 238000009413 insulation Methods 0.000 title claims abstract description 61
- 230000015556 catabolic process Effects 0.000 title claims abstract description 12
- 238000006731 degradation reaction Methods 0.000 title claims abstract description 9
- 238000000605 extraction Methods 0.000 title claims abstract description 9
- 230000007547 defect Effects 0.000 claims abstract description 57
- 239000007789 gas Substances 0.000 claims abstract description 25
- 238000013528 artificial neural network Methods 0.000 claims abstract description 15
- 238000012360 testing method Methods 0.000 claims abstract description 11
- 238000000034 method Methods 0.000 claims description 18
- 230000006866 deterioration Effects 0.000 claims description 5
- 238000004519 manufacturing process Methods 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 4
- LSJNBGSOIVSBBR-UHFFFAOYSA-N thionyl fluoride Chemical compound FS(F)=O LSJNBGSOIVSBBR-UHFFFAOYSA-N 0.000 description 17
- OBTWBSRJZRCYQV-UHFFFAOYSA-N sulfuryl difluoride Chemical compound FS(F)(=O)=O OBTWBSRJZRCYQV-UHFFFAOYSA-N 0.000 description 16
- 238000001514 detection method Methods 0.000 description 11
- 238000013101 initial test Methods 0.000 description 3
- 239000012212 insulator Substances 0.000 description 3
- 239000004020 conductor Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- QIVUCLWGARAQIO-OLIXTKCUSA-N (3s)-n-[(3s,5s,6r)-6-methyl-2-oxo-1-(2,2,2-trifluoroethyl)-5-(2,3,6-trifluorophenyl)piperidin-3-yl]-2-oxospiro[1h-pyrrolo[2,3-b]pyridine-3,6'-5,7-dihydrocyclopenta[b]pyridine]-3'-carboxamide Chemical class C1([C@H]2[C@H](N(C(=O)[C@@H](NC(=O)C=3C=C4C[C@]5(CC4=NC=3)C3=CC=CN=C3NC5=O)C2)CC(F)(F)F)C)=C(F)C=CC(F)=C1F QIVUCLWGARAQIO-OLIXTKCUSA-N 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 239000002923 metal particle Substances 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000002945 steepest descent method Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The GIS partial discharge characteristic parameter extraction method considering insulation degradation comprises the following steps: step 1, arranging SF oppositely by adopting a stage boosting mode6Carrying out 2 times of partial discharge repeatability tests on typical GIS insulation defects in gas; step 2: constructing a characteristic parameter set Q consisting of concentration ratios of different gases; and step 3: performing feature selection on feature parameters in the feature parameter set Q by using an mRMR algorithm to obtain priority sequence of the feature parameters, and sequentially selecting the first j feature parameters according to the sequence to reconstruct l sets Qj(ii) a And 4, step 4: and constructing an insulation defect identification model by using a back propagation neural network to obtain an optimal partial discharge characteristic parameter set. The invention can obtain the simplified and effective characteristic parameter combination for identifying the insulation defect and the more accurate identification effect.
Description
Technical Field
The invention relates to the technical field of electrician detection, in particular to a GIS (gas-insulated switch gear) partial discharge characteristic parameter extraction method considering insulation degradation.
Technical Field
Due to SF6The gas has excellent insulating and arc extinguishing performance, and the GIS is widely applied in the power industry. However, in the process of manufacturing, transporting, installing, overhauling, operating and the like of the GIS, some insulation defects inevitably occur inside the GIS, the defects can be gradually degraded in the long-term operation process, and partial discharge can occur inside equipment when the defects reach a certain degree. Partial discharge is a characteristic quantity which effectively represents the insulation condition of equipment, fault identification and severity degree evaluation are carried out by utilizing the partial discharge characteristic of the equipment, and the type of insulation defects and the partial discharge degree in the GIS can be mastered to a great extent. Therefore, the detection of partial discharge has important practical significance for ensuring the safe and reliable operation of the GIS.
The existing GIS partial discharge detection method mainly comprises the following steps: pulsed current methods, ultrasonic methods, optical detection methods, and ultrahigh frequency methods. The pulse current method is easy to be interfered by electromagnetic, and cannot be used for GIS field detection due to a GIS multipoint grounding structure; the ultrasonic method, the field equipment vibration, the corona noise and the like have great influence on the accuracy of ultrasonic detection; the optical detection method has low sensitivity, has detection dead angles, and cannot detect the partial discharge phenomenon which is not in the detection range of the photoelectric sensor; the ultrahigh frequency method is tedious in field discharge amount calibration and is easily subjected to random narrow-band interference the same as a detection frequency band.
Disclosure of Invention
The invention aims to provide a GIS partial discharge characteristic parameter extraction method considering insulation degradation, which can obtain a simplified and effective characteristic parameter combination and a more accurate identification effect for GIS insulation defect identification.
In order to achieve the purpose, the GIS partial discharge characteristic parameter extraction method considering insulation degradation is characterized by comprising the following steps:
step 1: adopts a step boosting mode to arrange in SF6Performing 2 times of partial discharge repeatability tests on typical GIS insulation defects in gas, wherein the tests are from the beginning of defect generation stable partial discharge to the ending of defect generation breakdown;
step 2: before the voltage of each stage is finished, acquiring and analyzing the types of gas insulation media in the GIS and the concentrations corresponding to the types, and constructing a characteristic parameter set Q consisting of concentration ratios of different gases by using the acquired types of the gas insulation media in the GIS and the concentration data corresponding to the types;
and step 3: utilizing an mRMR (maximum correlation minimum redundancy) algorithm to carry out feature sorting on feature parameters in a feature parameter set Q, searching feature parameters which have maximum correlation with the insulation defect types and minimum redundancy among the feature parameters from the feature parameter set Q, obtaining priority sorting of the feature parameters according to the searching result, and sequentially selecting j (j is 1,2, …, l, l is the number of elements in the feature parameter set Q) feature parameters to reconstruct l sets QjThat is, the result of sorting the l characteristic parameters in Q is: q. q.s1、q2…qlThen, the reconstructed l sets are respectively: q1＝{q1}；Q2＝{q1、q2}；…；Ql＝{q1、q2…ql}；
And 4, step 4: using Back Propagation Neural Networks (BPNN)n Neural Network) to construct an insulation defect recognition model, and sequentially collecting QjThe characteristic parameter (concentration ratio between different gases) in (j 1,2, …, l) is input into BP neural network as input quantity to obtain defect identification rate TjWhen T isjCharacteristic parameter set Q tending to be stablejNamely, the optimal partial discharge characteristic parameter set is obtained.
Typical GIS insulation defects in the step 1 of the technical scheme are insulation defects of the GIS in the manufacturing, transporting, installing, overhauling and running processes of the GIS.
In step 2 of the above technical solution, various gas insulating media A1、A2、…、AkIs represented by C (A)1)、C(A2)、…、C(Ak) And k represents the number of gas species generated in the test, and the set of characteristic parameters Q ═ C (a) is constructedh)/C(Ah+1),C(Ah)/C(Ah+2),…,C(Ah)/C(Ak) And (h represents a sequence number, h is 1,2, …, k-1), and the number of elements l in the set Q is k (k-1)/2.
In step 3 of the above technical solution, the mRMR algorithm selects the feature quantity according to the maximum statistical dependency principle, that is, m features having the maximum correlation with the insulation defect type and the minimum redundancy among them are searched from the feature set, and the correlation D and the redundancy R are defined as follows:
in the formula, Q and | Q | represent a feature set and the number of features included in the feature set, respectively, and I (x)i(ii) a c) Representing features x in a feature set QiAnd insulation defect class c, I (x)i；xj) Representing features x in a feature set QiAnd feature xjD (Q, c) represents the feature set Q and the insulation gapCorrelation between the trap classes c, r (Q) denotes the redundancy of the feature set Q; given two continuous random variables x and y with probability densities p (x) and p (y), respectively, and a joint probability density p (x, y), the mutual information between x and y can be defined as:
when x and y are discrete random variables, the above equation can be written as:
when the mRMR is used for feature selection, the correlation D between features and classes and the redundancy R inside the features are required to have the largest difference:
maxΦ1(D,R),Φ1＝D-R
assume that the set of all n features is QnFrom Q already according to the mRMR criterionnM features are selected, and the subset of the features is QmThe subset of remaining features is { Q }n-QmTo obtain Q }m+1From { Qn-QmFind the m +1 th feature in the sequence, and make it and QmCombined to form Qm+1Still satisfying the mRMR principle, then the m +1 th feature xiIt should satisfy:
the characteristic parameters are subjected to priority ordering by using the formula, namely the characteristic parameters meeting the formula are sequentially selected from the set Q, and the sequence of the selected characteristic parameters is the priority ordering; reconstructing the l sets QjIt is constructed according to the priority ranking of the characteristic parameters.
In step 4 of the above technical solution, the back propagation neural network parameters are set as: respectively using characteristic parameter set Qj(j1,2, …,10) as an input sample, taking the typical GIS insulation defect type in step 1 as an expected output of the neural network, and setting the allowable minimum deviation between the actual output and the expected output to be 0.01, the maximum iteration number to be 100 and the number of hidden layer neurons to be 10.
The invention applies test voltage to different insulation defects by adopting a stage boosting mode to consider SF6The method comprises the steps of utilizing an mRMR algorithm to construct a characteristic parameter set related to a gas concentration ratio, utilizing BPNN to construct an insulation defect identification model, and conducting training iteration on the constructed characteristic parameter set to obtain a simplified and effective characteristic parameter combination and a more accurate identification effect for GIS insulation defect identification.
The invention adopts a chemical detection method, is not influenced by environmental noise, electromagnetic interference and the like, and has accurate fault diagnosis.
Drawings
FIG. 1 shows the gas concentrations measured before the end of the initial test voltage duration.
Fig. 2 is a BPNN learning flowchart.
Fig. 3 shows BPNN identification results of 4 insulation defects.
In FIG. 2, IkThe sample data is referred to, X is for counting, the parameter initialization X ═ 0 indicates that the training sample has not started, X indicates the total number of samples, and X +1 > X indicates that all samples have been trained. Epsilon represents the minimum deviation allowed between the actual output and the desired output of the neural network.
Detailed Description
The invention is described in further detail below with reference to the following figures and specific examples:
the invention aims to provide a GIS partial discharge characteristic parameter extraction method considering insulation degradation. During the operation of GIS, SF is caused by insulation defect and different pressurizing time6Different degrees of deterioration of the gas-insulating medium, different SF6The partial discharge of the deteriorated GIS inevitably varies to some extent.
The GIS partial discharge characteristic parameter extraction method considering insulation degradation comprises the following steps:
step 1: adopts a step boosting mode to arrange in SF6Performing local discharge repeatability tests on typical GIS insulation defects in gas for 2 times, wherein the tests are from the beginning of stable local discharge of the defects to the end of breakdown of the defects, the step boosting amplitude is set as d, and the step voltage duration is set as e;
step 2: collecting and analyzing the type A of the gas insulation medium in the GIS before the voltage duration of each stage is overiAnd concentration C (A) thereofi) Constructing a characteristic parameter set Q by using the acquired data;
and step 3: performing feature selection on feature parameters in the set Q by using an mRMR algorithm to obtain priority sequence of the feature parameters, and sequentially selecting the first j feature parameters according to the sequence to reconstruct l sets Qj；
And 4, step 4: constructing an insulation defect identification model by using a Back Propagation Neural Network (BPNN), and sequentially collecting QjThe characteristic parameter (concentration ratio between different gases) in (j 1,2, …, l) is input into BP neural network as input quantity to obtain defect identification rate TjWhen T isjCharacteristic parameter set Q tending to be stablejNamely, the optimal partial discharge characteristic parameter set is obtained.
Further, the typical insulation defect of the GIS in step 1 refers to an insulation defect generated during the manufacturing, transporting, installing, overhauling, running and the like of the GIS. This example includes 4 insulation defect categories: the high-voltage insulator comprises abnormal metal protruding from the high-voltage conductor (called protrusion defect for short), metal particles or debris freely moving in a cavity (called particle defect for short), various dirt attached to the surface of the insulator (called dirt defect for short), and tiny air gaps formed between the high-voltage conductor and the basin-type insulator (called air gap defect for short).
Further, specific values of the initial test voltage, the end test voltage, the stage boosting amplitude d and the stage voltage duration e of the 4 defects in the step 1 are shown in the following table:
further, the type A of the gas insulation medium in the step 2iThe method comprises the following steps: CF (compact flash)4、CO2、SO2F2、SOF2、SO2The concentrations C (A) of these 5 gases were measured before the end of the initial test voltage durationi) The specific values of (A) are shown in FIG. 1.
Further, the characteristic parameter set Q in step 2 refers to a set formed by concentration ratios between different gases. The set of characteristic parameters Q ═ C (CF) constructed in this example4)/C(CO2)、C(CF4)/C(SO2F2)、C(CF4)/C(SOF2)、C(CF4)/C(SO2)、C(CO2)/C(SO2F2)、C(CO2)/C(SOF2)、C(CO2)/C(SO2)、C(SO2F2)/C(SOF2)、C(SO2F2)/C(SO2)、C(SOF2)/C(SO2) And f, the number l of elements in the set Q is 10.
Further, the mRMR algorithm in step 3 is a mutual information-based feature selection method, which selects the feature quantity according to the maximum statistical dependency principle, that is, m features having the maximum correlation with the insulation defect type and the minimum redundancy among each other are found from the feature set, and the correlation D and the redundancy R are defined as follows:
in the formula, Q and | Q | represent a feature set and the number of features included in the feature set, respectively, and I (x)i(ii) a c) Representing features x in a feature set QiAnd insulation defect classc mutual information between, I (x)i；xj) Representing features x in a feature set QiAnd feature xjD (Q, c) represents the correlation between the feature set Q and the insulation defect class c, and r (Q) represents the redundancy of the feature set Q; given two continuous random variables x and y with probability densities p (x) and p (y), respectively, and a joint probability density p (x, y), the mutual information between x and y can be defined as:
when x and y are discrete random variables, the above equation can be written as:
when the mRMR is used for feature selection, the correlation D between features and classes and the redundancy R inside the features are required to have the largest difference:
maxΦ1(D,R),Φ1＝D-R
assume that the set of all n features is QnFrom Q already according to the mRMR criterionnM features are selected, and the subset of the features is QmThe subset of remaining features is { Q }n-QmTo obtain Q }m+1From { Qn-QmFind the m +1 th feature in the sequence, and make it and QmCombined to form Qm+1Still satisfying the mRMR principle, then the m +1 th feature xiIt should satisfy:
further, the priority ranking of the feature parameters in the set Q obtained by using the mRMR algorithm in step 3 is: c (SO)2F2)/C(SOF2)、C(CF4)/C(CO2)、C(CF4)/C(SO2)、C(SOF2)/C(SO2)、C(CO2)/C(SO2F2)、C(CO2)/C(SO2)、C(SO2F2)/C(SO2)、C(CO2)/C(SOF2)、C(CF4)/C(SO2F2)，C(CF4)/C(SOF2)。
Further, the l (l ═ 10) sets Q reconstructed in step 3 arejSequentially comprises the following steps:
Q1＝{C(SO2F2)/C(SOF2)}；
Q2＝{C(SO2F2)/C(SOF2)、C(CF4)/C(CO2)}；
Q3＝{C(SO2F2)/C(SOF2)、C(CF4)/C(CO2)、C(CF4)/C(SO2)}；
Q10＝{C(SO2F2)/C(SOF2)、C(CF4)/C(CO2)、C(CF4)/C(SO2)、C(SOF2)/C(SO2)、C(CO2)/C(SO2F2)、C(CO2)/C(SO2)、C(SO2F2)/C(SO2)、C(CO2)/C(SOF2)、C(CF4)/C(SO2F2)，C(CF4)/C(SOF2)}。
furthermore, the BPNN in step 4 is a multi-layer feedforward neural network that performs network training based on an error back propagation algorithm, and the learning rule is to use a steepest descent method to continuously adjust the weight and the threshold of the network through back propagation, so that the sum of squares of the errors of the network is continuously reduced. The learning process of the BPNN is composed of forward propagation and backward propagation, the forward propagation is used for network computation, the output of a certain input is solved, the backward propagation is used for transmitting errors layer by layer, a connection weight and a threshold are modified, and the learning process is shown in fig. 2.
Further, for the BPNN parameter in step 4, the following settings are set: respectively using characteristic parameter set QjAnd (j ═ 1,2, …,10) concentration ratio data are used as input samples, 4 defect types are used as expected outputs of the neural network, and the minimum deviation epsilon allowed between the actual output and the expected output is set to be 0.01, the maximum iteration number is set to be 100, and the number of hidden layer neurons is set to be 10.
Further, the BPNN identification results for 4 insulation defects obtained in step 4 are shown in fig. 3, and it is apparent that when Q is setjWhen the number j of the characteristic parameters is more than or equal to 3, the identification accuracy rate exceeds 90 percent and basically tends to be stable, and in practical engineering application, in order to reduce redundancy, the set Q can be selected3The characteristic parameters in the step (2) are used as the optimal characteristic quantity combination for GIS insulation defect identification. In this example, Q is selected3＝{C(SO2F2)/C(SOF2)、C(CF4)/C(CO2)、C(CF4)/C(SO2) And the GIS insulation defect identification is carried out by taking the GIS insulation defect identification as a characteristic parameter, the identification accuracy rate is 93.75%, and the identification effect is good.
Details not described in this specification are within the skill of the art that are well known to those skilled in the art.
Claims (5)
1. A GIS partial discharge characteristic parameter extraction method considering insulation degradation is characterized by comprising the following steps:
step 1: adopts a step boosting mode to arrange in SF6Performing 2 times of partial discharge repeatability tests on typical GIS insulation defects in gas, wherein the tests are from the beginning of defect generation stable partial discharge to the ending of defect generation breakdown;
step 2: before the voltage of each stage is finished, acquiring and analyzing the types of gas insulation media in the GIS and the concentrations corresponding to the types, and constructing a characteristic parameter set Q consisting of concentration ratios of different gases by using the acquired types of the gas insulation media in the GIS and the concentration data corresponding to the types;
and step 3: performing characteristic parameter matching on characteristic parameters in the characteristic parameter set Q by using an mRMR algorithmLine feature sorting, searching feature parameters which have the maximum correlation with the insulation defect types and have the minimum redundancy among the feature parameters from a feature parameter set Q, obtaining the priority sorting of the feature parameters according to the searching result, sequentially selecting the front j (j is 1,2, …, l, l is the number of elements in the feature parameter set Q) feature parameters according to the sorting to reconstruct l sets QjThat is, the result of sorting the l characteristic parameters in Q is: q. q.s1、q2…qlThen, the reconstructed l sets are respectively: q1＝{q1}；Q2＝{q1、q2}；…；Ql＝{q1、q2…ql}；
And 4, step 4: constructing an insulation defect identification model by utilizing a back propagation neural network, and sequentially using a set QjThe characteristic parameter (concentration ratio between different gases) in (j 1,2, …, l) is input into BP neural network as input quantity to obtain defect identification rate TjWhen T isjCharacteristic parameter set Q tending to be stablejNamely, the optimal partial discharge characteristic parameter set is obtained.
2. The method of extracting GIS partial discharge characteristic parameters considering insulation deterioration according to claim 1, wherein: typical GIS insulation defects in the step 1 are insulation defects of the GIS in the manufacturing, transporting, installing, overhauling and running processes of the GIS.
3. The method of extracting GIS partial discharge characteristic parameters considering insulation deterioration according to claim 1, wherein: in step 2, various gas insulating media A1、A2、…、AkIs represented by C (A)1)、C(A2)、…、C(Ak) And k represents the number of gas species generated in the test, and the set of characteristic parameters Q ═ C (a) is constructedh)/C(Ah+1),C(Ah)/C(Ah+2),…,C(Ah)/C(Ak) And (h represents a sequence number, h is 1,2, …, k-1), and the number of elements l in the set Q is k (k-1)/2.
4. The method of extracting GIS partial discharge characteristic parameters considering insulation deterioration according to claim 1, wherein: in step 3, the mRMR algorithm selects the feature quantity according to the maximum statistical dependency principle, that is, m features having the maximum correlation with the insulation defect type and the minimum redundancy among the m features are searched from the feature set, and the correlation D and the redundancy R are defined as follows:
in the formula, Q and | Q | represent a feature set and the number of features included in the feature set, respectively, and I (x)i(ii) a c) Representing features x in a feature set QiAnd insulation defect class c, I (x)i；xj) Representing features x in a feature set QiAnd feature xjD (Q, c) represents the correlation between the feature set Q and the insulation defect class c, and r (Q) represents the redundancy of the feature set Q; given two continuous random variables x and y with probability densities p (x) and p (y), respectively, and a joint probability density p (x, y), the mutual information between x and y can be defined as:
when x and y are discrete random variables, the above equation can be written as:
when the mRMR is used for feature selection, the correlation D between features and classes and the redundancy R inside the features are required to have the largest difference:
maxΦ1(D,R),Φ1＝D-R
assume that the set of all n features is QnFrom Q already according to the mRMR criterionnM features are selected, and the subset of the features is QmThe subset of remaining features is { Q }n-QmTo obtain Q }m+1From { Qn-QmFind the m +1 th feature in the sequence, and make it and QmCombined to form Qm+1Still satisfying the mRMR principle, then the m +1 th feature xiIt should satisfy:
the characteristic parameters are subjected to priority ordering by using the formula, namely the characteristic parameters meeting the formula are sequentially selected from the set Q, and the sequence of the selected characteristic parameters is the priority ordering; reconstructing the l sets QjIt is constructed according to the priority ranking of the characteristic parameters.
5. The method of extracting GIS partial discharge characteristic parameters considering insulation deterioration according to claim 1, wherein: in the step 4, the back propagation neural network parameters are set as: respectively using characteristic parameter set QjAnd (j-1, 2, …,10) using the concentration ratio data as an input sample, using the typical GIS insulation defect type in the step 1 as the expected output of the neural network, and setting the allowable minimum deviation between the actual output and the expected output to be 0.01, the maximum iteration number to be 100 and the number of hidden layer neurons to be 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011226040.XA CN112485610B (en) | 2020-11-05 | 2020-11-05 | GIS partial discharge characteristic parameter extraction method considering insulation degradation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011226040.XA CN112485610B (en) | 2020-11-05 | 2020-11-05 | GIS partial discharge characteristic parameter extraction method considering insulation degradation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112485610A true CN112485610A (en) | 2021-03-12 |
| CN112485610B CN112485610B (en) | 2024-02-02 |
Family
ID=74928724
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011226040.XA Active CN112485610B (en) | 2020-11-05 | 2020-11-05 | GIS partial discharge characteristic parameter extraction method considering insulation degradation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112485610B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113568353A (en) * | 2021-07-30 | 2021-10-29 | 国网四川省电力公司电力科学研究院 | Switch cabinet partial discharge monitoring and early warning method based on reduced state feature set |
| CN113607271A (en) * | 2021-07-15 | 2021-11-05 | 国网电力科学研究院武汉南瑞有限责任公司 | GIL defect online monitoring system and method based on vibration signals |
| CN113640629A (en) * | 2021-07-26 | 2021-11-12 | 国网电力科学研究院武汉南瑞有限责任公司 | GIS partial discharge state evaluation method, recording medium and system |
| CN115048985A (en) * | 2022-05-17 | 2022-09-13 | 国网浙江省电力有限公司嘉兴供电公司 | Electrical equipment fault discrimination method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107656177A (en) * | 2017-08-18 | 2018-02-02 | 西安交通大学 | Based on shelf depreciation or SF6The dry sleeve insulation defect diagnostic method of analyte |
| CN108120907A (en) * | 2018-01-10 | 2018-06-05 | 清华大学 | The partial discharge diagnostic method of feature extraction under a kind of tremendously low frequency voltage based on power frequency |
| CN109683066A (en) * | 2018-11-08 | 2019-04-26 | 山东康威通信技术股份有限公司 | Power cable typical defect local discharge signal recognition methods |
| CN109871746A (en) * | 2018-12-30 | 2019-06-11 | 国网辽宁省电力有限公司电力科学研究院 | A kind of mode identification method based on the diagnosis of GIS insulation defect |
-
2020
- 2020-11-05 CN CN202011226040.XA patent/CN112485610B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107656177A (en) * | 2017-08-18 | 2018-02-02 | 西安交通大学 | Based on shelf depreciation or SF6The dry sleeve insulation defect diagnostic method of analyte |
| CN108120907A (en) * | 2018-01-10 | 2018-06-05 | 清华大学 | The partial discharge diagnostic method of feature extraction under a kind of tremendously low frequency voltage based on power frequency |
| CN109683066A (en) * | 2018-11-08 | 2019-04-26 | 山东康威通信技术股份有限公司 | Power cable typical defect local discharge signal recognition methods |
| CN109871746A (en) * | 2018-12-30 | 2019-06-11 | 国网辽宁省电力有限公司电力科学研究院 | A kind of mode identification method based on the diagnosis of GIS insulation defect |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113607271A (en) * | 2021-07-15 | 2021-11-05 | 国网电力科学研究院武汉南瑞有限责任公司 | GIL defect online monitoring system and method based on vibration signals |
| WO2023284127A1 (en) * | 2021-07-15 | 2023-01-19 | 国网电力科学研究院武汉南瑞有限责任公司 | Vibration-signal-based online monitoring system for gil defects, and method |
| CN113640629A (en) * | 2021-07-26 | 2021-11-12 | 国网电力科学研究院武汉南瑞有限责任公司 | GIS partial discharge state evaluation method, recording medium and system |
| CN113568353A (en) * | 2021-07-30 | 2021-10-29 | 国网四川省电力公司电力科学研究院 | Switch cabinet partial discharge monitoring and early warning method based on reduced state feature set |
| CN115048985A (en) * | 2022-05-17 | 2022-09-13 | 国网浙江省电力有限公司嘉兴供电公司 | Electrical equipment fault discrimination method |
| CN115048985B (en) * | 2022-05-17 | 2024-02-13 | 国网浙江省电力有限公司嘉兴供电公司 | Electrical equipment fault discrimination method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112485610B (en) | 2024-02-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112485610A (en) | GIS partial discharge characteristic parameter extraction method considering insulation degradation | |
| US11002789B2 (en) | Analog circuit fault feature extraction method based on parameter random distribution neighbor embedding winner-take-all method | |
| CN108761287B (en) | Transformer partial discharge type identification method | |
| Jineeth et al. | Classification of partial discharge sources in XLPE cables by artificial neural networks and support vector machine | |
| CN116682458A (en) | GIS partial discharge voiceprint detection method based on improved wavelet packet of energy operator | |
| CN115015719A (en) | Sleeve partial discharge recognition algorithm based on optical fiber ultrasonic sensing signal | |
| CN109557434B (en) | Partial discharge signal identification method under strong background noise based on composite dictionary sparse representation classification | |
| CN114397569A (en) | Circuit breaker fault arc detection method based on VMD parameter optimization and sample entropy | |
| CN113740671A (en) | Fault arc identification method based on VMD and ELM | |
| CN116415182A (en) | Arc additive low-carbon steel fatigue crack prediction method based on causality and schematic injection force | |
| Evagorou et al. | Evaluation of partial discharge denoising using the wavelet packets transform as a preprocessing step for classification | |
| Yufeng et al. | Partial discharge pattern recognition of DC XLPE cables based on convolutional neural network | |
| CN113963757A (en) | Oil-filled electrical equipment fault diagnosis method based on gas relation and graph neural network | |
| KR20220043548A (en) | Partial-discharge detecting method and device | |
| Florentino et al. | Spectral Subband Centroid Energy Vectors Algorithm and Artificial Neural Networks for Acoustic Emission Pattern lassification. | |
| Ortego et al. | Deep Learning Tools Analysis for Automatic Partial Discharge Detection Based on PRPD Patterns | |
| Evagorou et al. | Performance of the Support Vector Machine Partial Discharge classification method to noise contamination using phase synchronous measurements | |
| Xiao et al. | The two-level fault diagnosis model of SF 6 electrical equipment | |
| Backe et al. | Efficient feature extraction for machine-learning-based PD classification in medium voltage cables | |
| Yeo et al. | Predicting of partial discharge in medium-voltage cables using recurrent neural network with long short-term memory cells after wavelet denoising | |
| KR102592309B1 (en) | Automatic decision method using partial discharge pulse of underground transmission cable | |
| Abdullah et al. | MODELLING OF PARTIAL DISCHARGE ANALYSIS SYSTEM USING WAVELET TRANSFORM DENOISING TECHNIQUE IN LABVIEW ENVIRONMENT | |
| CN113419141B (en) | Direct-current line fault positioning method and device | |
| CN117607624A (en) | Fault detection method for power distribution equipment based on acoustic signals | |
| Wang et al. | Partial Discharge Signal Denoising Based on Block Thresholding |
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 |