CN112420136B - Latent fault tracing method for sulfur hexafluoride high-voltage equipment - Google Patents
Latent fault tracing method for sulfur hexafluoride high-voltage equipment Download PDFInfo
- Publication number
- CN112420136B CN112420136B CN202011359734.0A CN202011359734A CN112420136B CN 112420136 B CN112420136 B CN 112420136B CN 202011359734 A CN202011359734 A CN 202011359734A CN 112420136 B CN112420136 B CN 112420136B
- Authority
- CN
- China
- Prior art keywords
- concentration
- sulfur hexafluoride
- gas
- voltage equipment
- hexafluoride high
- 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.)
- Active
Links
- 229910018503 SF6 Inorganic materials 0.000 title claims abstract description 46
- SFZCNBIFKDRMGX-UHFFFAOYSA-N sulfur hexafluoride Chemical compound FS(F)(F)(F)(F)F SFZCNBIFKDRMGX-UHFFFAOYSA-N 0.000 title claims abstract description 46
- 229960000909 sulfur hexafluoride Drugs 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 39
- 239000007789 gas Substances 0.000 claims abstract description 114
- 238000007621 cluster analysis Methods 0.000 claims abstract description 32
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 claims abstract description 16
- 229910052799 carbon Inorganic materials 0.000 claims abstract description 16
- 239000011159 matrix material Substances 0.000 claims description 7
- 239000002131 composite material Substances 0.000 claims description 6
- 238000000556 factor analysis Methods 0.000 claims description 4
- 238000000528 statistical test Methods 0.000 claims description 3
- 238000010998 test method Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 abstract description 4
- 238000000513 principal component analysis Methods 0.000 description 18
- 230000008569 process Effects 0.000 description 8
- 238000011160 research Methods 0.000 description 8
- 239000000126 substance Substances 0.000 description 8
- 238000006243 chemical reaction Methods 0.000 description 6
- 239000012535 impurity Substances 0.000 description 5
- 238000000354 decomposition reaction Methods 0.000 description 4
- 239000004033 plastic Substances 0.000 description 4
- 230000009467 reduction Effects 0.000 description 4
- 239000004575 stone Substances 0.000 description 4
- 239000011797 cavity material Substances 0.000 description 3
- 230000001186 cumulative effect Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 239000000047 product Substances 0.000 description 3
- 238000007619 statistical method Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 229910052760 oxygen Inorganic materials 0.000 description 2
- 239000001301 oxygen Substances 0.000 description 2
- 238000007789 sealing Methods 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 240000002853 Nelumbo nucifera Species 0.000 description 1
- 235000006508 Nelumbo nucifera Nutrition 0.000 description 1
- 235000006510 Nelumbo pentapetala Nutrition 0.000 description 1
- -1 Polytetrafluoroethylene Polymers 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000007795 chemical reaction product Substances 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 229920001343 polytetrafluoroethylene Polymers 0.000 description 1
- 239000004810 polytetrafluoroethylene Substances 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 229940126680 traditional chinese medicines Drugs 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/30—Prediction of properties of chemical compounds, compositions or mixtures
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Computing Systems (AREA)
- Chemical & Material Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Crystallography & Structural Chemistry (AREA)
- Water Supply & Treatment (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention relates to a latent fault tracing method of sulfur hexafluoride high-voltage equipment, which comprises the steps of firstly detecting the concentration of each trace gas in a cavity of the equipment to form standardized index data, classifying the trace gas into 10 classes through cluster analysis, analyzing the association degree between the classified trace gas and variables before the cluster analysis, classifying the gas into 5 classes, and finally carrying out statistical inspection on the standardized index data to obtain the threshold concentration of each gas, H 2 Or CH (CH) 4 Exceeding the threshold concentration by invasive, CF 4 、C 2 F 6 、CS 2 Or C 3 F 8 The concentration exceeding the threshold is common carbon type, CO 2 、SO 2 F 2 、SO 2 COS or CO exceeding a critical concentration to severe carbon, H 2 S exceeds a threshold concentration and is medium, CS 2 The concentration exceeding the threshold is mixed; when the gas exceeding the threshold concentration contains more than two kinds of 5 kinds of trace gases, the gas isCorresponding to the superposition of fault types.
Description
Technical Field
The invention relates to the technical field of principal component analysis and cluster analysis, in particular to a method for tracing latent faults of sulfur hexafluoride high-voltage equipment.
Background
SF 6 Equipment such as high voltage circuit breakers can age gradually during operation, with latent failures often associated with chemical reactions in extreme environments, accompanied by SF 6 Polytetrafluoroethylene and impurity gas O as equipment cavity material 2 H and H 2 The complex reaction of O, the type and content of the reaction products are important for researching the chemical reaction in the black box, and the SF with faults is obtained 6 The high-voltage electric equipment has found at least 13 other trace gases, the possible fault types can be explored through the information reflected by the messenger gases, and the messenger gases are the gases carrying chemical reaction information, so that the design and materials of the high-voltage equipment are improved, and the performance of the equipment is improved; obtainingThe safety information of the equipment is used and maintained, and the service life of the equipment is prolonged.
Several methods have been developed to obtain SF 6 SF removal in high voltage circuit breaker and other equipment 6 Besides information on other gases due to extreme conditions, students are also trying to directly correlate these gas information with latent faults of such devices. However, due to SF 6 The black box nature of the space where the high-voltage circuit breaker and other devices fail, the exploration of the chemical events actually occurring in the devices is still in the preliminary stage, meanwhile, the combination of the types and concentration levels of the messenger gases obtained presents diversity and complexity, and the rules which can directly and definitely relate the messenger gases to the related chemical events are very limited at present; in this respect, it is conventional practice to correlate the level of single messenger gas in a single case with a certain possible chemical event, which requires researchers to be very familiar and specialized with high pressure equipment processes, materials, chemical bases, etc., is not suitable for general research, and is obviously inefficient and difficult to deal with when faced with the large amount of complex information provided by actual conditions. Thus, there is a strong need for a reliable and efficient method of correlating chemical information contained in a large number of messenger gases with their possible failure processes.
The social science statistical software package (English abbreviation SPSS) is a program system for counting and analyzing various types of data, and has powerful functions of Principal Component Analysis (PCA) and Cluster Analysis (CA). CA refers to an analysis process of grouping a collection of physical or abstract objects into a plurality of classes according to some characteristics which are not yet clear, and in the classification process, a classification standard is not required to be given in advance, and a clustering structure is found through cluster analysis so as to realize automatic data aggregation, thereby realizing exploration of a large number of data commonalities. PCA and CA have been used in some fields, and PCA is used in medicine to explore the rule of symptom distribution of a large number of patients, and PCA and CA are used in medicine to explore the effective components of traditional Chinese medicines. In the electric power field, lotus groups and the like establish a wind farm equivalent modeling method based on PCA and CA, and the method provides a reliable means for analyzing the safety and stability of a power system after wind power grid connection.
It can be seen from theory and practice that for SPSS, the powerful statistical function of CA is applied in many fields and therefore can be applied to SF 6 The reliable and efficient latent fault type association analysis is performed by equipment such as a high-voltage circuit breaker, but no relevant report exists at present.
Disclosure of Invention
For SF 6 The invention provides a sulfur hexafluoride high-voltage equipment latent fault tracing method, which selects a large number of SF with latent faults 6 SF in high voltage circuit breaker 6 The protection gas is used as a research object, the type and content level of the specific micro messenger gas in the protection gas are used as monitoring indexes, the classification of faults is explored through cluster analysis, and the application display is well supported from the principle of chemical reaction by adopting the conclusion obtained by the method.
The invention is realized by the following technical scheme:
a method for tracing latent faults of sulfur hexafluoride high-voltage equipment comprises the following steps:
when H in normalized index data 2 Or CH (CH) 4 The latent fault type of the sulfur hexafluoride high-voltage equipment is invasive when the concentration exceeds the respective threshold concentration; when CF in normalized index data 4 、C 2 F 6 、、CS 2 Or C 3 F 8 When the concentration of the sulfur hexafluoride high-voltage equipment exceeds the respective threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is common carbon type; when the standard index data contains CO 2 、SO 2 F 2 、SO 2 When the concentration of COS or CO exceeds the respective threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is serious carbon; when H in normalized index data 2 When the concentration of S exceeds the threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is medium type; when CS is in standardized index data 2 When the concentration of the sulfur hexafluoride high-voltage equipment exceeds the threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is mixed;
when the trace gas exceeding the threshold concentration in the standardized index data contains more than two types of the 5 types of trace gas in the step 3, the latent fault type of the sulfur hexafluoride high-voltage equipment is composite, and the composite is superposition of the corresponding fault types of each trace gas exceeding the threshold concentration.
Preferably, the number of sulfur hexafluoride high-voltage devices in the step 1 is several, and the sulfur hexafluoride high-voltage devices are from the same working condition.
Further, the number of sulfur hexafluoride high-voltage devices in the step 1 is not less than 40.
Preferably, step 3 analyzes the degree of association between the trace gas after classification and the variable before the cluster analysis by taking the similarity of the variables at the time of 3 kinds of variable classification, 4 kinds of variable classification and 5 kinds of variable classification as factors, respectively.
Further, step 3 performs single factor analysis of variance on the 3 variable classifications, the 4 variable classifications, and the 5 variable classifications, respectively, and classifies all trace gases into 5 classes.
Preferably, step 4 obtains the threshold concentration of each gas as follows;
the outlier of each gas concentration in the standardized index data is removed through a Q test method, and the concentration average value and the concentration average deviation of each gas are calculated, wherein the threshold concentration of each gas=the concentration average value of each gas+4×the concentration average deviation of each gas.
Preferably, H in step 4 2 Is 0.8ppm, CO is 36ppm, CH 4 Is 30ppm, CF 4 Is 84ppm and CO 2 Is 12ppm, C 2 F 6 Is 66ppm and SO 2 F 2 Is 1.1ppm, H 2 S has a threshold concentration of 0.25ppm and C 3 F 8 Is 24ppm, COS is 2.3ppm, SOF 2 Is 2.4ppm and SO 2 Is 20ppm and CS 2 Is 2.6ppm.
Compared with the prior art, the invention has the following beneficial technical effects:
the invention relates to sulfur hexafluoride high-voltage equipmentThe method for tracing the latent faults comprises the steps of firstly detecting the concentration of each trace gas in a sulfur hexafluoride high-voltage equipment cavity with the latent faults so as to form standardized index data, classifying the standardized index data by taking the concentration content of all trace gases as variables through cluster analysis, classifying all trace gases into 10 types, further analyzing the association degree between the classified trace gases and the variables before the cluster analysis, classifying all trace gases into 5 types, and finally classifying the latent faults of the sulfur hexafluoride high-voltage equipment into invasive types, common carbon types, serious carbon types, medium types, mixed types or superposition of the above types by combining the threshold concentration and classification results of each gas. The invention adopts a large number of SF with latent faults 6 The information of messenger gas in the high-voltage circuit breaker is used as a monitoring index, a rule hidden in a large amount of information is discovered by using a statistical method, blindness and uncertainty in the process of researching a few cases are greatly reduced, the selection of the messenger gas is included in the research results of the existing independent cases, and possible chemical principles are considered; SF (sulfur hexafluoride) 6 The high-voltage circuit breaker latent fault occurs in a closed space, the process of the high-voltage circuit breaker latent fault has black box attribute, and cluster analysis in statistical analysis can be classified according to the characteristics which are not clear yet, and then the classified standard and the main variables influencing classification are reversely tracked, so that the high-voltage circuit breaker latent fault is particularly suitable for the research of the black box process. The invention establishes an SF based on cluster analysis 6 The method for tracing the faults of the high-voltage equipment can reflect the chemical essence behind a large number of faults by utilizing the statistical advantages, and classifies the same faults and discovers the common characteristics of the faults by reducing the dimension of the inspected gas information. Analysis of information containing specific 13 messenger gases in gases associated with latent faults by this method found that 3 of the 13 messenger gases could be the marker information variable, while the chemical characteristics of these 3 main variables and the final variable H of the cluster analysis 2 、CF 4 CO and CO 2 The characteristics are highly consistent, and fault types are classified into invasive types, common carbon types, severe carbon types, medium types, mixed types and the like. The method is to obtain a large amount of SF 6 High-voltage equipment fault tracing provides oneAn efficient method is a large amount of SF 6 The statistical classification of latent faults of high-voltage electrical equipment and messenger gas confirmation provide an effective means, also provide direction guidance for individual case tracing, can be parallel and complemented with independent case research in application, synchronously update research results along with application breadth and depth, and is SF 6 The system safety and stability of the high-voltage electric product provide a new control method for SF in the high-voltage transmission industry 6 The analysis of latent faults and safe operation of the equipment provide an efficient tool.
Drawings
FIG. 1 is a lithotripsy diagram of the principal component analysis according to the present invention.
FIG. 2 is a graph of the gas variable spectrum according to the present invention.
Detailed Description
The invention will now be described in further detail with reference to specific examples, which are intended to illustrate, but not to limit, the invention.
The invention relates to a method for tracing latent faults of sulfur hexafluoride high-voltage equipment, which uses SF through principal component analysis and cluster analysis of SPSS 6 50 detected high-voltage circuit breakers contain 13 trace impurity gases (H 2 、CO、CH 4 、CF 4 、CO 2 、C 2 F 6 、SO 2 F 2 、H 2 S、C 3 F 8 、COS、SOF 2 、SO 2 、CS 2 ) SF of (2) 6 For example, a gas sample is subjected to statistical analysis based on various trace gas contents, main impurity gases with indication characteristics are searched, and the feasibility of classifying the sample is further summarized and related to potential fault classification of the high-voltage switch.
Generally, 13 trace impurity gases are the smallest amount; the gas content information of the above samples is shown in table 1. However, it should be noted that the data in Table 1 are from 50 SF under the same conditions 6 High voltage circuit breaker, each SF 6 The high-voltage circuit breaker is collected for 3 times, and the average value is obtained after detection.
TABLE 1 gas content information (ppm) for 50 samples
The present invention also provides a validation of CA by analyzing the above samples with PCA in SPSS to determine those gases or combinations thereof that are more relevant to potential failure.
PCA is also a method used mathematically to reduce dimensions by orthogonally transforming a set of variables that may have correlations into a set of new, linearly uncorrelated complex variables, often accompanied by the reduction of dimensions of statistical variables, by selecting, as needed, information methods where fewer complex variables reflect as many of the original variables as possible, the reduction of variables meaning a reduction in computation.
The main component analysis mainly comprises the following steps: (1) determining the number m of principal components; (2) Principal component F i An expression.
Step (1): determination of m of the number of principal Components
Firstly, the standardized index data of table 1 are recorded in PCA of SPSS, the number of samples forming the standardized index data can be less than 50 or more than 50, then variable views are sequentially switched, analysis is carried out, dimension reduction is selected, factor analysis is carried out, a description dialog box is set, initial solution and KMO and Buttery sphericity test are selected, characteristic values corresponding to 13 new variables are obtained, as shown in table 2, and accordingly the determination process of the number m of main components is started.
TABLE 2 principal component extraction analysis Table
The eigenvalue is an index representing the magnitude of the influence of a new variable, if it is greater than 1, it means that it is relative toThe interpretation dynamics of the new variable is larger than the average interpretation dynamics of 1 original variable, so that the main component can be firstly included; as can be seen from Table 2, the principal component analysis has feature values corresponding to the first 3 new variables arranged according to the feature values of each of them greater than 1, respectively lambda 1 ＝6.93，λ 2 ＝3.29，λ 3 ＝2.15。
The above procedure can be summarized as follows: and converting standardized index data formed by all trace impurity gases in sulfur hexafluoride high-voltage equipment with latent faults into a corresponding number of new variables through principal component analysis, wherein each gas corresponds to one new variable, analyzing characteristic values corresponding to the new variables, and finally taking the new variables with the characteristic values larger than 1 into principal components to obtain the number of the principal components.
The contribution of the first 3 new variables to the variance is 43.5%, 28.8% and 22.9%, respectively, the cumulative variance contribution of the first 3 new variables reaches 95.2% >80.0%, because 80.0% is the minimum requirement of a reasonable principal component selection result to the principal component cumulative variance contribution, so that the initial selection of the first 3 new variables to be incorporated into the principal component is correct, because not all new variables will have a cumulative variance contribution greater than 80.0%.
And then selecting to present a crushed stone map, wherein the crushed stone map is obtained by the characteristic value being larger than 1 and the iteration number being 25, and is shown in fig. 1, and the crushed stone map is drawn according to the interpretation degree of the difference of each new variable to data, wherein the abscissa of the crushed stone map represents the number of the main components, and the ordinate represents the characteristic value. In general, a new variable corresponding to a steeper part of the lithotriptic graph curve is selected as a main component, a corresponding lithotriptic graph is obtained from the data in table 1, and it can be seen from fig. 1 that the characteristic value after the 4 th new variable tends to be flat, which suggests that the number of the preferred main components is 3.
From the analysis, it can be considered that the first 3 new variables cover most of the information of the first 13 variables, so that gas information and SF are performed 6 The analysis of the correlation of the latent fault types of the high-voltage circuit breaker can be simplified to consider only the 3 main components, thus the linear group based on the original 13 variablesThe combination forms 3 new variables, thereby reducing the dimension and difficulty of calculation. Extracting the first 3 new variables as main components, and marking the main component factors as F respectively 1 、F 2 And F 3 。
Correlation analysis between principal component and primary variable
When the correlation degree between the principal component and the original variable is examined by adopting the principal component analysis function of the SPSS, the matrix is automatically rotated in the PCA, and then the maximum variance is selected, which is performed simultaneously with the obtaining of table 2, so that the intra-group distance is the shortest and the inter-group distance is the longest, and the correlation result between the principal component and the original variable can be obtained, as shown in table 3.
TABLE 3 load matrix
It can be seen that: SO (SO) 2 F 2 、CO 2 、SOF 2 、SO 2 The 6 primary variables of COS and CO have higher load on the 1 st main component, which shows that the 1 st main component basically reflects the information of the 6 primary variables, and similarly the 2 nd main component basically reflects C 2 F 6 、C 3 F 8 、CF 4 、CS 2 Information of the 4 original variables, the 3 rd principal component basically reflects H 2 、CH 4 、H 2 S information of these 3 kinds of original variables.
Step (2): principal component Fi expression
Expression F of 3 principal components has not been directly obtained from tables 2 and 3 i Each load in the initial factor load matrix in table 3 represents the correlation coefficient between the principal component and the corresponding primary variable, and the weighting coefficient of the primary variable can be obtained by dividing the load by the eigenvalue of the principal component corresponding to table 2 and then opening. Finally obtain F i The expression is as follows: f (F) 1 ＝0.35ZX 1 +0.35ZX 2 +0.34ZX 3 +0.34ZX 4 +0.34ZX 5 +0.45ZX 6 +0.26ZX 7 +0.22ZX 9 +0.22ZX 10 +0.22ZX 11 +0.20X 12 +0.22ZX 13 ；F 2 ＝0.53ZX 8 +0.44ZX 9 +0.45ZX 10 +0.44ZX 11 ；F 3 ＝0.45ZX 7 +0.56ZX 12 +0.55ZX 13 The method comprises the steps of carrying out a first treatment on the surface of the Wherein Z is a pending vector; xi is the original variable, i.e. the content of the ith gas of the corresponding sample in Table 1, i.e. X 1 Is H 2 Content of X 2 For the content of CO, X 3 Is CH 4 And so on, according to the formula f=λ 1 F 1 /(λ 1 +λ 2 +λ 3 )+λ 2 F 2 /(λ 1 +λ 2 +λ 3 )+λ 3 F 3 /(λ 1 +λ 2 +λ 3 ) With normalized ZX i Performing multiple linear regression analysis for dependent variables, and establishing a regression model to obtain a final expression: f=0.12 ZX 1 +0.20ZX 2 +0.19ZX 3 +0.19ZX 4 +0.19ZX 5 +0.25ZX 6 +0.23ZX 7 +0.14ZX 8 +0.24ZX 9 +0.22ZX 10 +0.22ZX 11 +0.20ZX 12 +0.21ZX 13 。
Principal component analysis results
By performing principal component analysis on 13 kinds of fault gases, it can be concluded from the final expression: ZX (ZX) 6 、ZX 7 、ZX 9 、ZX 10 、ZX 11 、ZX 13 The previous coefficients are greater than 0.2, indicating the variation of these gas concentrations and SF 6 The faults of the high-voltage circuit breaker are greatly related, and when C exists 2 F 6 、SO 2 F 2 、C 3 F 8 、COS、SOF 2 And CS (common services) 2 SF when the concentration of the gases is changed 6 High voltage circuit breakers have failed to varying degrees.
The invention determines those gases or combinations thereof that are more relevant to potential faults by performing cluster analysis on a large number of samples. The cluster analysis mainly determines the number of variables (i.e., the number of clusters) n.
The normalized index data formed in table 1 is categorized. Firstly, standardized index data of table 1 are recorded in CA of SPSS, then variable views are sequentially switched, analysis is performed, systematic clustering is selected, cluster analysis is performed, individual cases are moved into individual case mark data, statistics is selected, and an approximation matrix is performed, so that table 4 is obtained. The approximation is used to represent the distance between an individual and another individual in the subclass, and the closer the data is to 1, the closer the distance between an individual and another individual in the subclass is, so that the two individuals are considered to be replaced by each other, and thus, the clusters can be included; and selecting the range of the solution to obtain different classification sets, selecting single factor variance analysis for different groups, selecting the classification with the greatest contribution to the clustering result, analyzing and selecting Pearson correlation coefficients of variables, respectively merging with the Pearson correlation coefficients being more than or equal to 0.99, more than or equal to 0.9 and more than or equal to 0.8 to obtain a table 5, selecting square Euclidean distance in the method, and selecting the direction of a graph to be horizontal to obtain a pedigree graph shown in fig. 2.
TABLE 4 variable approximation matrix
TABLE 5 associated gas merging results
As can be seen from tables 4 and 5, when the Pearson correlation coefficient is 0.99 or more, H 2 And CH (CH) 4 Can replace each other, CF 4 And C 2 F 6 、C 3 F 8 Can replace each other, C 2 F 6 、C 3 F 8 Can replace each other, remove the replaced gas, and finally select 10The gas-like body is used as a research object. When the Pearson correlation coefficient is 0.93 or more, the kind of the selected gas is 5 kinds. When the Pearson correlation coefficient is more than or equal to 0.8, the selected gas is 3 kinds, so that the research efficiency is greatly improved.
The extraction of clusters can also be derived from a graph of graphs, plotted against distance between variables, where the abscissa represents distance and the ordinate represents gas type, which can be used to help determine the optimal number of clusters. The shorter the distance, the smaller the difference of each gas, so that the earlier the gas is polymerized into a class; the longer the distance, the greater the variability, the later the group.
In summary, the first 13 variables can be classified into 3 categories, so that gas information and SF are performed 6 In the analysis of the correlation of the latent fault types of the high-voltage circuit breaker, in order to reduce the dimension and difficulty of calculation, the method can be simplified to select one of the 3 types respectively. Extracting 3 variables as clustering results, wherein the clustering factors are Z respectively 1 、Z 2 、Z 3 . Cluster analysis was performed on this basis.
Cluster optimization based on the selected variable categories is as follows:
when the correlation degree between clusters and original variables is examined by adopting the cluster analysis function of SPSS, standardized index data of table 1 is input, the following operations in CA are sequentially selected, analysis, comparison average value and single factor variance test are sequentially carried out, all gas contents are taken as variables, samples formed by 10 types of gases shown in table 5 are classified, variable similarity (9 groups in total) in classification of 3 types, 4 types and 5 types of variables is sequentially taken as factors for analysis, the range of limiting classification is 3-5 types, and the results are collated as shown in table 6.
It can be seen that: the single-factor analysis of variance is performed on different classes, so that the significance of classifying the faults into 5 classes is best under the selection of three correlation coefficients, and therefore the faults are determined to be classified into 5 classes.
Table 6 significance levels of different variables after three classifications of variables and three correlation coefficient substitutions
Then for 50 SF 6 The raw data of the high-voltage circuit breaker, namely table 1, is subjected to statistical test, and the specific steps are as follows: and removing outliers of the concentration of each gas by a Q test method, and calculating to obtain a concentration average value and a concentration average deviation of each gas, wherein the threshold concentration of each gas = the concentration average value of each gas +4 x the concentration average deviation of each gas.
H 2 Is 0.8ppm, CO is 36ppm, CH 4 Is 30ppm, CF 4 Is 84ppm and CO 2 Is 12ppm, C 2 F 6 Is 66ppm and SO 2 F 2 Is 1.1ppm, H 2 S has a threshold concentration of 0.25ppm and C 3 F 8 Is 24ppm, COS is 2.3ppm, SOF 2 Is 2.4ppm and SO 2 Is 20ppm and CS 2 Is 2.6ppm.
The threshold concentration of each gas forms a threshold beyond which a fault is predicted. The following explanation of fault type classification is made in connection with experimental variables, where the chemical reaction mechanism has been considered:
(1)H 2 type, invasive failure, often defined by H 2 O intrusion is caused.
(2)CF 4 Type, common carbon type failures are often caused by decomposition of plastic parts.
(3)CO 2 Type, severe carbon type failure, high risk, often accompanied by component seal failure.
(4)H 2 S-type, dielectric type failures, often accompanied by H 2 O intrusion and plastic part decomposition.
(5)CS 2 Type, hybrid faults, which are the most damaging ones, are often associated with a number of causes.
When the trace gas exceeding the threshold concentration in the standardized index data contains more than two types of the 5 types of trace gas in the step 3, the latent fault type of the sulfur hexafluoride high-voltage equipment is composite, and the composite is superposition of the corresponding fault types of each trace gas exceeding the threshold concentration.
Generally, the severity of the fault types is in the order of a hybrid fault>Dielectric failure>Severe carbon failure>Common carbon type failure>An intrusive fault. Such as: the judgment basis of the invasive fault is H 2 With CS 2 If H is present in the system 2 Then is H 2 O intrusion. If no oxygen is introduced into the system, part S is CS 2 Exists in a form; if oxygen is intruded, CS 2 Will not exist in this form, its existing form is CO 2 Instead of it. CS is therefore 2 Can be used as O 2 Criteria for intrusion.
Application 1: the certain ultra-high voltage transformer station has tripping faults, and the SF in the tripping faults is tested by researchers 6 The data of Table 7 were obtained by measuring the decomposed gas and the humidity.
TABLE 7 switch 1 gas decomposition products and humidity test
As can be seen from Table 7, SO of group C 2 And H 2 S concentration is obviously higher than that of the A and B groups, and the C group shows serious carbon type faults and medium type faults, and has sealing failure and O 2 、H 2 O intrusion breaks down the plastic part. According to the results of the cover opening inspection and the disassembly inspection, O is used for disassembling scratch in transportation 2 、H 2 O enters the circuit breaker, consistent with the model predicted outcome.
Application 2: h is detected by an air chamber of 220kV transformer substation brake air chamber of 7 months of 2018 2 S、SO 2 Table 8 shows that the contents of the three gases, CO, are abnormal.
TABLE 8 detection data for switch 2 decomposition products
As can be seen from Table 8, CO and H 2 S concentration is high, severe carbon type faults and medium type faults are shown, and sealing failure and O exist 2 、H 2 O intrusion breaks down the plastic part. The results of the ultrasonic detection and the can opening experiment inspection are consistent with the prediction of the model.
Claims (7)
1. The method for tracing the latent fault of the sulfur hexafluoride high-voltage equipment is characterized by comprising the following steps:
step 1, detecting the concentration of each trace gas in a cavity of sulfur hexafluoride high-voltage equipment with latent faults, wherein the trace gas comprises H 2 、CO、CH 4 、CF 4 、CO 2 、C 2 F 6 、SO 2 F 2 、H 2 S、C 3 F 8 、COS、SOF 2 、SO 2 And CS (common services) 2 Wherein each gas corresponds to one concentration data to form standardized index data;
step 2, classifying the standardized index data by using the concentration and the content of all the trace gases as variables through cluster analysis to obtain a variable approximation matrix, and classifying all the trace gases into 10 types by taking the Pearson correlation coefficient as a standard and taking the Pearson correlation coefficient as a standard, wherein the Pearson correlation coefficient is greater than or equal to 0.99; wherein H is 2 And CH (CH) 4 Of the first class, CO of the first class, CF 4 、C 2 F 6 And C 3 F 8 To one kind, CO 2 To one kind, SO 2 F 2 To be of the same kind, H 2 S is a kind, COS is a kind, SOF 2 To one kind, SO 2 To one kind, CS 2 Is one type;
step 3, analyzing the association degree between the classified trace gases and the variables before cluster analysis on the basis of classifying all the trace gases into 10 types, and further classifying all the trace gases into 5 types; wherein H is 2 And CH (CH) 4 To one kind, CF 4 、C 2 F 6 、SOF 2 And C 3 F 8 To one kind, CO 2 、SO 2 F 2 、SO 2 COS and CO are one class, H 2 S is a kind, CS 2 Is one ofClass;
step 4, firstly, carrying out statistical test on the standardized index data to obtain the threshold concentration of each gas, and then combining the classification result of the step 3 to obtain the following conclusion so as to finish tracing the latent fault of the sulfur hexafluoride high-voltage equipment;
when H in normalized index data 2 Or CH (CH) 4 The latent fault type of the sulfur hexafluoride high-voltage equipment is invasive when the concentration exceeds the respective threshold concentration; when CF in normalized index data 4 、C 2 F 6 、CS 2 Or C 3 F 8 When the concentration of the sulfur hexafluoride high-voltage equipment exceeds the respective threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is common carbon type; when the standard index data contains CO 2 、SO 2 F 2 、SO 2 When the concentration of COS or CO exceeds the respective threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is serious carbon; when H in normalized index data 2 When the concentration of S exceeds the threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is medium type; when CS is in standardized index data 2 When the concentration of the sulfur hexafluoride high-voltage equipment exceeds the threshold concentration, the latent fault type of the sulfur hexafluoride high-voltage equipment is mixed;
when the trace gas exceeding the threshold concentration in the standardized index data contains more than two types of the 5 types of trace gas in the step 3, the latent fault type of the sulfur hexafluoride high-voltage equipment is composite, and the composite is superposition of the corresponding fault types of each trace gas exceeding the threshold concentration.
2. The method for tracing latent faults of sulfur hexafluoride high voltage equipment according to claim 1, wherein the number of sulfur hexafluoride high voltage equipment in the step 1 is several, and the sulfur hexafluoride high voltage equipment is from the same working condition.
3. The method for tracing latent faults of sulfur hexafluoride high voltage equipment according to claim 2, wherein the number of sulfur hexafluoride high voltage equipment in the step 1 is not less than 40.
4. The method for tracing latent faults of sulfur hexafluoride high-voltage equipment according to claim 1, wherein step 3 is characterized in that the degree of association between the classified trace gas and the variables before cluster analysis is analyzed by taking the similarity of the variables in 3 variable classifications, 4 variable classifications and 5 variable classifications as factors respectively.
5. The method for tracing latent faults of sulfur hexafluoride high voltage equipment according to claim 4, wherein step 3 carries out single factor analysis of variance on 3 variable classifications, 4 variable classifications and 5 variable classifications respectively, and all trace gases are classified into 5 classes.
6. The method for tracing latent faults of sulfur hexafluoride high voltage equipment according to claim 1, wherein step 4 obtains the threshold concentration of each gas as follows;
the outlier of each gas concentration in the standardized index data is removed through a Q test method, and the concentration average value and the concentration average deviation of each gas are calculated, wherein the threshold concentration of each gas=the concentration average value of each gas+4×the concentration average deviation of each gas.
7. The method for tracing latent fault of sulfur hexafluoride high voltage equipment according to claim 1, wherein in step 4, H 2 Is 0.8ppm, CO is 36ppm, CH 4 Is 30ppm, CF 4 Is 84ppm and CO 2 Is 12ppm, C 2 F 6 Is 66ppm and SO 2 F 2 Is 1.1ppm, H 2 S has a threshold concentration of 0.25ppm and C 3 F 8 Is 24ppm, COS is 2.3ppm, SOF 2 Is 2.4ppm and SO 2 Is 20ppm and CS 2 Is 2.6ppm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011359734.0A CN112420136B (en) | 2020-11-27 | 2020-11-27 | Latent fault tracing method for sulfur hexafluoride high-voltage equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011359734.0A CN112420136B (en) | 2020-11-27 | 2020-11-27 | Latent fault tracing method for sulfur hexafluoride high-voltage equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112420136A CN112420136A (en) | 2021-02-26 |
| CN112420136B true CN112420136B (en) | 2023-05-23 |
Family
ID=74842916
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011359734.0A Active CN112420136B (en) | 2020-11-27 | 2020-11-27 | Latent fault tracing method for sulfur hexafluoride high-voltage equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112420136B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113588886A (en) * | 2021-08-09 | 2021-11-02 | 国网河北省电力有限公司电力科学研究院 | Sulfur hexafluoride gas full-flow cycle monitoring system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH05256747A (en) * | 1992-03-12 | 1993-10-05 | Hitachi Ltd | Apparatus and method for diagnosing abnormality of electric apparatus containing perfluorocarbon |
| CN104793085A (en) * | 2014-07-16 | 2015-07-22 | 胡小青 | Novel distribution transformer control method of IEC method |
| CN108983052A (en) * | 2018-07-26 | 2018-12-11 | 武汉大学 | One kind being based on SF6The three classes failure triangle diagnostic method of air insulating device decomposition components |
| CN109508729A (en) * | 2018-09-21 | 2019-03-22 | 国网重庆市电力公司电力科学研究院 | A kind of POF method for diagnosing faults based on clustering |
| CN109900831A (en) * | 2019-04-04 | 2019-06-18 | 国网陕西省电力公司电力科学研究院 | Trace amounts of CO and CO in a kind of high concentrations of gas2Quantitative analysis method |
| CN110726785A (en) * | 2019-10-25 | 2020-01-24 | 国网陕西省电力公司电力科学研究院 | SF analysis based on GC-Q-ToF-MS6Method for medium trace permanent gas |
-
2020
- 2020-11-27 CN CN202011359734.0A patent/CN112420136B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH05256747A (en) * | 1992-03-12 | 1993-10-05 | Hitachi Ltd | Apparatus and method for diagnosing abnormality of electric apparatus containing perfluorocarbon |
| CN104793085A (en) * | 2014-07-16 | 2015-07-22 | 胡小青 | Novel distribution transformer control method of IEC method |
| CN108983052A (en) * | 2018-07-26 | 2018-12-11 | 武汉大学 | One kind being based on SF6The three classes failure triangle diagnostic method of air insulating device decomposition components |
| CN109508729A (en) * | 2018-09-21 | 2019-03-22 | 国网重庆市电力公司电力科学研究院 | A kind of POF method for diagnosing faults based on clustering |
| CN109900831A (en) * | 2019-04-04 | 2019-06-18 | 国网陕西省电力公司电力科学研究院 | Trace amounts of CO and CO in a kind of high concentrations of gas2Quantitative analysis method |
| CN110726785A (en) * | 2019-10-25 | 2020-01-24 | 国网陕西省电力公司电力科学研究院 | SF analysis based on GC-Q-ToF-MS6Method for medium trace permanent gas |
Non-Patent Citations (5)
| Title |
|---|
| Artificial Intelligence-Based Techniques for Emerging Heterogeneous Network: State of the Arts, Opportunities, and Challenges;XIAOFEI WANG等;《SPECIAL SECTION ON ARTIFICIAL INTELLIGENCE ENABLED NETWORKING》;20150828;第1379-1391页 * |
| Multivariate Evaluation Method for Screening Optimum Gas-Sensitive Materials for Detecting SF6 Decomposition Products;Jifeng Chu等;《ACS Sensors》;20200701;第1-24页 * |
| SF6 断路器绝缘介质和导电回路常见故障的处理探讨;杨凯迪;《科技成果纵横》;20200228;第80页 * |
| 六氟化硫中痕量一氧化碳和二氧化碳的气相色谱检测及其在断路器故障分析中的应用;杨韧等;《色谱》;20200630;第38卷(第6期);第702-707页 * |
| 基于改进型主成分分析的电力变压器潜伏性故障诊断;杨廷方等;《电力自动化设备》;20150630(第06期);第149-153页+165页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112420136A (en) | 2021-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107301296B (en) | Data-based qualitative analysis method for circuit breaker fault influence factors | |
| Lin et al. | Nonlinear dynamic principal component analysis for on-line process monitoring and diagnosis | |
| CN111598150B (en) | Transformer fault diagnosis method considering operation state grade | |
| CN113962259B (en) | Multi-mode double-layer fault diagnosis method for fuel cell system | |
| CN110297841B (en) | Transformer fault diagnosis and rapid indexing method and system | |
| Kirkbas et al. | Fault diagnosis of oil-immersed power transformers using common vector approach | |
| CN108304567B (en) | Method and system for identifying working condition mode and classifying data of high-voltage transformer | |
| CN104267668A (en) | Bayes-method-based spaceflight valve part fault diagnosis method in machining process | |
| CN112904810B (en) | Process industry nonlinear process monitoring method based on effective feature selection | |
| CN109389325B (en) | Method for evaluating state of electronic transformer of transformer substation based on wavelet neural network | |
| CN110794360A (en) | Method and system for predicting fault of intelligent electric energy meter based on machine learning | |
| CN108830006B (en) | Linear-nonlinear industrial process fault detection method based on linear evaluation factor | |
| CN112420136B (en) | Latent fault tracing method for sulfur hexafluoride high-voltage equipment | |
| Guo et al. | Fault diagnosis of power transformers based on comprehensive machine learning of dissolved gas analysis | |
| CN114091549A (en) | Equipment fault diagnosis method based on deep residual error network | |
| McGrail et al. | Data mining techniques to assess the condition of high voltage electrical plant | |
| CN115809407A (en) | Breaker fault diagnosis method and system based on vibration characteristic analysis | |
| Cheemala et al. | Transformer incipient fault diagnosis using machine learning classifiers | |
| CN114509620A (en) | Transformer fault diagnosis method based on monitoring data | |
| CN112085064B (en) | Transformer fault diagnosis method based on multi-classification probability output of support vector machine | |
| CN111458630B (en) | Fault diagnosis method for high-voltage circuit breaker | |
| CN111737993B (en) | Method for extracting equipment health state from fault defect text of power distribution network equipment | |
| CN115860243A (en) | Fault prediction method and system based on industrial Internet of things data | |
| Kaur et al. | Fault Diagnosis for Oil Immersed Transformer Using Certainty Factor | |
| CN112881827B (en) | Oil-immersed transformer fault diagnosis method based on improved grey correlation analysis |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20240523 Address after: No. 669, Hangtian Middle Road, Xi'an national civil aerospace industry base, Shaanxi 710100 Patentee after: Electric Power Research Institute of State Grid Shaanxi Electric Power Co.,Ltd. Country or region after: China Patentee after: State Grid Shaanxi Electric Power Co.,Ltd. Patentee after: STATE GRID CORPORATION OF CHINA Patentee after: National Network (Xi'an) Environmental Protection Technology Center Co.,Ltd. Address before: No.669, Hangtian Middle Road, Chang'an District, Xi'an City, Shaanxi Province Patentee before: STATE GRID SHAANXI ELECTRIC POWER Research Institute Country or region before: China Patentee before: STATE GRID SHAANXI ELECTRIC POWER Co. Patentee before: STATE GRID CORPORATION OF CHINA |