CN107633349A - Fault impact factor quantitative analysis method based on high-voltage switch gear - Google Patents
Fault impact factor quantitative analysis method based on high-voltage switch gear Download PDFInfo
- Publication number
- CN107633349A CN107633349A CN201710751628.9A CN201710751628A CN107633349A CN 107633349 A CN107633349 A CN 107633349A CN 201710751628 A CN201710751628 A CN 201710751628A CN 107633349 A CN107633349 A CN 107633349A
- Authority
- CN
- China
- Prior art keywords
- voltage switch
- switch gear
- quantitative analysis
- impact factor
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- 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
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The present invention provides the fault impact factor quantitative analysis method based on high-voltage switch gear, can obtain the precise relation between fault impact factor and corresponding fault type.Methods described includes：(1) the fault type data after sorting out according to operational parameter data and data obtain primary data；(2) primary data is standardized；(3) regression analysis is introduced as dependent variable using operational parameter data as independent variable, fault category data, obtains optimal high-voltage switch gear fault impact factor Quantitative Analysis Model；(4) regression diagnostics is carried out to the Logistic regression equations of model, it is qualified to enter (5)；It is unqualified, then suboptimum high-voltage switch gear fault impact factor Quantitative Analysis Model in (3) is taken, regression diagnostics is carried out again and enters (5) until qualified；(5) according to high-voltage switch gear fault impact factor Quantitative Analysis Model qualified in (4), high-voltage switch gear fault impact factor is obtained by its Logistic regression equation quantitative analysis.
Description
Technical field
The present invention relates to the analysis of the O＆M of high-voltage switch gear, the fault impact factor quantitative analysis specially based on high-voltage switch gear
Method.
Background technology
High-voltage switch gear refer to rated voltage 3kV and more than, be mainly used in cut-offfing and close the electrical equipment of galvanic circle.Its alias
It is called primary cut-out, such product can not only be cut off or no-load current and load current in closed high voltage circuit, and
When system jam by the effect of relay protection, overload electric current and short circuit current are cut off, it has suitable
Perfect arc extinguishing structure and enough cutout abilities.High-voltage switch gear is usually mechanical-electrical-hydraulic integration equipment, and it is main in power network
In the different substation of each department, play and cut-off the normal power transmission and distribution of circuit realiration, and protection is played in overload line disconnection
Effect.It is big with usage amount, it is distributed the features such as scattered.Due to these features, such product is easily sent out during operation of power networks
Raw all kinds of failures, can generally cause large area blackout, not only influence resident living, and industrial production is caused damage.
High-voltage switch gear, due to the generation of various failures, the operation/maintenance data note of correlation is had in unit of operation during O＆M
Record.The operation/maintenance data mainly recorded comprising " fault category " and during a series of generation of failures it is related to equipment operation because
Element.Under normal circumstances, the operation/maintenance data is limited only to record and accumulation, and sufficiently analysis is not carried out to it and is dug
Pick, i.e., the relation not being found between " fault category " that contains the inside and operation correlative factor.Current failure influence factor
Discovery mainly obtained by operating personnel or technical staff by operation/maintenance data by its experience, in face of increasingly sophisticated
During fault impact factor analysis problem intelligent, automation high-voltage switch gear is more hidden, it appears increasingly there is subjective limitation
Property, it is difficult to the maintenance work of reality instruct and corresponding reference frame comprehensively.
The content of the invention
For problems of the prior art, the present invention provides a kind of fault impact factor based on high-voltage switch gear and quantified
Analysis method, quantitative analysis is carried out, the precise relation between fault impact factor and corresponding fault type can be obtained, be follow-up
Maintenance work accurate foundation and reference are provided, greatly improve the service life of high-voltage switch gear, reduce fault rate.
The present invention is to be achieved through the following technical solutions：
Fault impact factor quantitative analysis method based on high-voltage switch gear, comprises the following steps,
Step 1, the failure operation/maintenance data of high-voltage switch gear to be analyzed is obtained by each Utilities Electric Co.'s information system；Described event
Barrier operation/maintenance data comprises at least high-voltage switch gear to be analyzed and the fault type data of nature of trouble is recorded when breaking down, and treats point
Analysis high-voltage switch gear records the operational parameter data of O＆M information when breaking down；After sorting out according to operational parameter data and data
Fault type data obtain primary data sample；
Step 2, primary data sample is pre-processed as follows；
Step 2.1, two-category data resource is formed with fault type data, two-category data resource is mapped, will
Fault type is each mapped to 0 or 1；
Step 2.2, the discrete data in primary data sample is mapped as numeric type；
Step 2.3, on the basis of the processing of step 2.1 and step 2.2, primary data sample is standardized,
Obtain the unified normal data sample of dimension；
Step 3, based on Binary Logistic regression analysis algorithm, according to obtained normal data sample, operation is joined
Number data introduce regression analysis as independent variable, and fault category data introduce regression analysis as dependent variable, establish high-voltage switch gear
Fault impact factor Quantitative Analysis Model；Tested by test data set, preferably obtain optimal high-voltage switch gear fault impact
Factor Quantitative Analysis Model；
Step 4, the Logistic regression equations of optimal high-voltage switch gear fault impact factor Quantitative Analysis Model are returned
Return diagnosis, if regression diagnostics is qualified to enter step 5；If unqualified, need to take suboptimum high-voltage switch gear failure shadow in step 3
The factor of sound Quantitative Analysis Model, carries out regression diagnostics to its regression equation again, enters step 5 until regression diagnostics is qualified；
Step 5, according to the qualified high-voltage switch gear fault impact factor Quantitative Analysis Model of regression diagnostics in step 4, by it
Logistic regression equation quantitative analyses obtain high-voltage switch gear fault impact factor.
Preferably, in step 2.3, primary data sample is standardized by min-max standardized method,
The codomain of data in primary data sample is mapped in [0,1].
Further, the processing of min-max standardization is carried out to primary data sample according to following formula,
Wherein, x* represents the original variable in primary data sample, and x represents the variable after standardization, and max represents former
The upper dividing value of beginning variable codomain, min represent the floor value of original variable codomain.
Preferably, the step of also including rejecting independent variable in step 3；
After high-voltage switch gear fault impact factor Quantitative Analysis Model is established, carry out conspicuousness using method of gradual regression and pick
Remove, the variable for not meeting AIC index criterions is rejected from regression equation, then the independent variable ultimately remained in regression equation is
Failure is occurred with the independent variable significantly affected.
Preferably, in step 3, tested by test data set, preferably obtain optimal high-voltage switch gear fault impact because
Plain Quantitative Analysis Model comprises the following steps that,
A. operational parameter data collection is randomly divided into size identical n parts, n is the positive integer more than 1, in each run
A copy of it is selected as inspection set, and remaining is all training set, and model construction is carried out by Logistic regression algorithms；
B. by step a Repeated ms time, n is the positive integer more than 1, as a result will obtain n m times grader, and by each
Training and inspection, the classification accuracy of each grader can be obtained；
C. in calculation procedure b m times of n grader average classification accuracy, computational methods are to take m times of n to classify
The average of device accuracy rate；Classification accuracy is found afterwards closest to the object classifiers of average classification accuracy；The target classification
Device is exactly the high-voltage switch gear fault impact factor Quantitative Analysis Model that model preferably obtains.
Preferably, in step 4, to the Logistic recurrence sides of optimal high-voltage switch gear fault impact factor Quantitative Analysis Model
Cheng Jinhang regression diagnosticses, wherein the index method diagnosed is likelihood ratio test method.
Preferably, in step 5, the Logistic recurrence sides of described high-voltage switch gear fault impact factor Quantitative Analysis Model
Cheng Wei：
G (x)=ω0+ω1x1+ω2x2+ω3x3 (1-2)
Wherein, x=(x1,x2,x3) represent that independent variable is the vector that each operational parameter data is formed, p (y=1 | x) represent therefore
Hinder the conditional probability occurred；ω0Represent the intercept of regression equation, ω1、ω2、ω3Independent variable x in regression equation is represented respectively1、
x2、x3Coefficient；
It is as follows so as to obtain the result of quantitative analysis,
Failure is made a difference by independent variable coefficient magnitude relation three variables of judgement, and degree is descending to put in order；
Respective fault impact factor is corresponded to according to each operational factor, using independent variable coefficient as weights, passes through weights
The quantization of size obtains the size for the influence degree that failure occurs each fault impact factor.
Further, in described step 5, each fault impact factor is obtained by the quantization of weights size failure is occurred
Influence degree size when, when independent variable coefficient is bigger, the influence degree that failure occurs is bigger.
Compared with prior art, the present invention has technique effect beneficial below：
(1) present invention is using Logistic recurrence as rudimentary algorithm, it is proposed that high-voltage switch gear fault impact factor quantitative analysis
Method.This method compares other quantitative analysis methods, can significantly more efficient searching high-voltage switch gear fault type and each O＆M because
Quantitative relationship between element, and adapt to data type in high-voltage switch gear operation/maintenance data and not only included continuous data but also comprising discrete
Type data, and fault type is as the situation that dependent variable is more classification discrete types.
(2) used in the present invention in order to meet the requirement of Logistic recurrence by fault data pretreatment by more points
The problem of class, is converted into the problem of two classification and is analyzed, and is returned compared to more classification Logistic are directly used in the demand analysis
Return analysis to simplify the degree of difficulty of analysis, and the analysis demand of the scene is not only suitable for by analyzing this method, analyze simultaneously
As a result easily explain and explanation.
(3) make use of n times in the present invention in order to ensure that fault impact factor Quantitative Analysis Model has certain stability
The method of m foldings, constructs some fault impact Factor Models, and have chosen optimum analysis model by certain method.
For only once being analyzed, there is obvious robustness.
It is idle with wasting present invention, avoiding operation/maintenance data, and provide corresponding ginseng for the operation and maintenance process of high-voltage switch gear
Examine foundation.It can analyze to obtain which of O＆M process hazards and specific fault is had a great influence, with the mathematics of science
Model reflects its influence factor, and reference frame is provided for high-voltage switch gear O＆M, can be more preferable beneficial to follow-up O＆M
Prevent and fix a breakdown, improve the use reliability and stability of high-voltage switch gear.
Brief description of the drawings
Fig. 1 is the FB(flow block) of method described in present example.
Embodiment
With reference to specific embodiment, the present invention is described in further detail, it is described be explanation of the invention and
It is not to limit.
The present invention can carry out quantitative excavate and analysis to the fault impact factor of high-voltage switch gear operation/maintenance data.Based on data
Mining algorithm, by the specific mathematical model of structure, the high-voltage switch gear operation/maintenance data for recording and accumulating is carried out going deep into excavation, obtained
Relation between fault type and part O＆M factor, a complete quantitative analysis system is formed, so as to be high-voltage switch gear
Equipment fault Quantitative Analysis of Influence Factors provides corresponding method.It is base by being pre-processed to high-voltage switch gear operation/maintenance data
Place mat is performed in the high-voltage switch gear quantitative analysis of operation/maintenance data, is quantitatively divided so as to improve high-tension switch gear fault impact factor
The result stability of analysis method, ensure the reliable results of data analysis.
High-voltage switch gear fault impact factor quantitative analysis method of the present invention, the quantitative analysis based on Logistic regression algorithms
Method is proposed, a set of more perfect analysis method is provided with making full use of for the excavation of high-voltage switch gear fault data.Solve
Operation/maintenance data idle the problem of fully excavating with utilizing, while form a set of fault impact factor quantitative analysis for a long time
Analysis system.
As shown in figure 1, the present invention comprises the following steps that:
Step (1)：Dependent failure operation/maintenance data is obtained by each Utilities Electric Co.'s information system, the failure operation/maintenance data is necessary
Comprising two kinds of data, a kind of is to the fault type data for recording nature of trouble, such as " actuating machine when high-voltage switch gear breaks down
Structure is abnormal " etc., another kind is the operational parameter data of the record O＆M information recorded when high-voltage switch gear breaks down, for example " throws
Transport the time ".
Step (2) high-voltage switch gear data prediction mainly includes the pretreatment of three parts.
Fault type packet classification containing various faults in 2.1 high-voltage switch gear operation/maintenance datas, and it is polytypic
Logistic regression analyses have certain difficulty in terms of result explanation, and polytypic problem is divided into two points in the present invention
The problem of class, i.e., by the processing in data plane, avoided in the fault impact factor quantitative analysis method of operation/maintenance data
Returned using polytypic Logistic, and use Binary Logistic regression, it is as a result explanatory stronger, it is more beneficial for failure
Influence factor is excavated.Specific practice is as follows：
2.11. two-category data resource is formed.Assuming that the fault type in operation/maintenance data in fault type data is F1、F2、
F3、F4、F5, wherein F1、F2、F3、F4For the clear and definite failure of property, and failure F5For the indefinite other failures of property.First by number
It is property clearly some failure and the indefinite other events of property to be broken according to collection as 4 data subsets, each data subset
Barrier combines to form, such as：F in failure1With failure F5Two fault datas have combined to form F1The fault impact factor analysis of failure
Data resource S1。
2.12. to two-category data resource impact.The data of each failure subset in step 2.11 are mapped, had
Body is for example：By failure F1Data resource S1In " fault type " be F11 is mapped as, represents failure, by " fault type "
It is F50 is mapped as, represents that failure does not occur.
2.2 operation/maintenance data discrete data mappings are handled.Because discrete data is introduced directly into Logistic recurrence sides
Cheng Zhonghui causes the setting of dummy variable, and explanation that so can be on fault impact Multi-factor analysis has influence.Therefore this is special
Discrete data is mapped as numeric type in profit.For example " the environment gradation for surface pollution " in operation/maintenance data in operational factor is shared
" a ", " b ", " c ", " d ", " e " this 5 grades.By grade from low to high be each mapped to " 1 ", " 2 ", " 3 ", " 4 ",
“5”。
The processing of 2.3 data normalizations.Data are not standardized before establishing Logistic regression models will
Have influence on the degree of accuracy of final mask.Used here as min-max standardized method standardization.This method can be by original
The codomain of beginning data is mapped in [0,1].For its processing mode as shown in formula 1-1, wherein x* represents original variable, and x represents to pass through
Variable after standardization, max represent the upper dividing value of original variable codomain, and min represents the floor value of original variable codomain.
Step (3) establishes high-voltage switch gear fault impact factor Quantitative Analysis Model.High-voltage switch gear fault impact factor quantifies
Analysis model is based on Binary Logistic regression.Why reason is returned based on Logistic and be high-voltage switch gear operation/maintenance data
In fault type data be classifying type variable, it is desirable to the quantitative pass established between fault type data and each fault impact factor
System, Logistic regression algorithms are adapted to the most.All fault impact factors to be analyzed are incorporated into back as independent variable first
Return in analysis, carry out conspicuousness rejecting using method of gradual regression afterwards, the variables of AIC index criterions will not met from regression equation
Middle rejecting, then it will be that failure is occurred with the variable significantly affected to ultimately remain in the variable in regression equation.
Because Logistic regression algorithms, its output is the result of two classification, can regard that sorting algorithm carries out model as
Training, and tested by test data set.It is proposed that rolling over cross validation using n m tries to achieve stable failure shadow in the present invention
The factor of sound Quantitative Analysis Model.Preferable n and m takes 10, carries out ten ten folding cross validations, and specific practice is：
A. the operation/maintenance data collection that the operation/maintenance data obtained in step 1 is formed is randomly divided into 10 parts of size identical, every
A copy of it is selected during secondary operation as inspection set, and remaining is all training set, and carried out by Logistic regression algorithms
Model construction.Here it is the way of ten foldings in so-called " ten ten foldings ".
B. it will be repeated ten times the step of a, be " ten ten foldings ".As a result 100 graders will be obtained, and by each
Training and inspection, the classification accuracy of each grader can be obtained.
C. the average classification accuracy of 100 graders in b is calculated, computational methods are the equal of 100 grader accuracys rate
Value.Classification accuracy is found afterwards closest to that grader of average classification accuracy.The grader is exactly that model selects excellent obtain
The high-voltage switch gear fault impact factor Quantitative Analysis Model arrived.
Step (4)：The regression equation of optimal Quantitative Analysis Model to being obtained in step (3) carries out regression diagnostics, if
Regression diagnostics is qualified to enter step (5), if unqualified, needs to take in step (3) suboptimum close to average classification accuracy
Grader, again to its regression equation carry out regression diagnostics.Qualified and its classification accuracy is diagnosed from average mark until finding
Untill the nearest grader of class accuracy rate.The index method wherein diagnosed is likelihood ratio test method.
Step (5)：To examining qualified Quantitative Analysis Model in step (4), from the Logistic regression equations in model
In obtain high-voltage switch gear fault impact factor.Assuming that high-voltage switch gear fault impact factor Quantitative Analysis Model is：
G (x)=ω0+ω1x1+ω2x2+ω3x3 (1-2)
Wherein x=(x1,x2,x3) to represent independent variable be the vector of each factors composition, p (y=1 | x) represents what failure occurred
Conditional probability.ω0Represent the intercept of regression equation, ω1、ω2、ω3Variable x in regression equation is represented respectively1、x2、x3Coefficient.
Assuming that ω in Quantitative Analysis Model1＞ ω2＞ ω3, three variables can be determined that by the magnitude relationship of coefficient inequality
The degree that made a difference to failure is descending to put in order as x1＞ x2＞ x3.Thus can obtain influenceing the shadow that failure occurs
The factor of sound, provides certain reference frame, maintenance work personnel can pay close attention on this basis for the O＆M of high-voltage switch gear
Related factor.
Below by taking the primary cut-out operation/maintenance data of somewhere as an example, the incidence relation between failure and influence factor is analyzed.
Used here as the DAS RStudio instruments under Windows environment, illustrate the specific of fault impact factor quantitative analysis
Flow and step.
(1) operation/maintenance data of primary cut-out is collected, field information is as shown in table 1 in final sample data, specific sample
The example of content is as shown in table 2.10 samples are only listed herein carries out example.
The sample data field information that table 1 is collected
The sample data example that table 2 is collected
(2) sample data is pre-processed, includes three parts：Form two-category data resource, to two-category data
Resource is mapped, data is standardized.So that " fault type " is operating mechanism exception as an example, the data of table 2 are entered
The sample data as shown in table 3, table 4 is obtained after line number Data preprocess.
The classification of table 3 two processing and the data resource after mapping
Data resource of the table 4 after standardization
(3) it is dependent variable to select " fault type " in table 4, the factor in addition to " fault type " as independent variable, with
Logistic establishes fault impact factor Quantitative Analysis Model based on returning.Ten ten folding cross validations have been used in this patent
Method, data have been randomly divided into 10 parts first, have taken 9 parts to be used as model training every time in turn, take 1 part as examine, altogether
Obtain 10 graders.By the way that said process is carried out ten times, 100 graders are finally given, calculate the classification of each grader
Accuracy rate, and it is 86% to obtain average classification accuracy, the grader for finding the closest value is the 35th grader.Pass through
Its likelihood ratio test is notable to be found to the regression diagnostics of the model, showing the equation of the regression model can use.
(4) by taking " operating mechanism abnormal " as an example, its optimum classifier is exactly regression model, that is, regression equation.By
The variable conspicuousness of method of gradual regression is rejected, and the Logistic regression equation results that finally give are：
First pass around variable conspicuousness reject after regression equation in variable represent on failure have significantly affect.Root
According to regression equation (1-4) it can be seen that " operating mechanism is abnormal " failure occurs and usage time (use_time) and drop-out current
Number (open_current), and year environment temperature (envir_temp) have obvious relation.And according to each variation coefficient
Size may determine that these three variables descending order of degree that made a difference to " operating mechanism abnormal " this failure is " to make
With the time ", " drop-out current number ", " environment temperature ".The fault impact factor of quantitative analysis is resulting in, is high-voltage switch gear
O＆M provide reference frame.
Claims (8)
1. the fault impact factor quantitative analysis method based on high-voltage switch gear, it is characterised in that comprise the following steps,
Step 1, the failure operation/maintenance data of high-voltage switch gear to be analyzed is obtained by each Utilities Electric Co.'s information system；Described failure fortune
Dimension data comprises at least the fault type data that nature of trouble is recorded when high-voltage switch gear to be analyzed breaks down, and height to be analyzed
Compress switch and the operational parameter data of O＆M information is recorded when breaking down；Failure after sorting out according to operational parameter data and data
Categorical data obtains primary data sample；
Step 2, primary data sample is pre-processed as follows；
Step 2.1, two-category data resource is formed with fault type data, two-category data resource is mapped, by failure
Type is each mapped to 0 or 1；
Step 2.2, the discrete data in primary data sample is mapped as numeric type；
Step 2.3, on the basis of the processing of step 2.1 and step 2.2, primary data sample is standardized, obtained
The unified normal data sample of dimension；
Step 3, based on Binary Logistic regression analysis algorithm, according to obtained normal data sample, by operational factor number
Regression analysis is introduced according to as independent variable, fault category data introduce regression analysis as dependent variable, establish high-voltage switch gear failure
Quantitative Analysis of Influence Factors model；Tested by test data set, preferably obtain optimal high-voltage switch gear fault impact factor
Quantitative Analysis Model；
Step 4, recurrence is carried out to the Logistic regression equations of optimal high-voltage switch gear fault impact factor Quantitative Analysis Model to examine
It is disconnected, if regression diagnostics is qualified to enter step 5；If unqualified, need to take in step 3 suboptimum high-voltage switch gear fault impact because
Plain Quantitative Analysis Model, regression diagnostics is carried out to its regression equation again, enter step 5 until regression diagnostics is qualified；
Step 5, according to the qualified high-voltage switch gear fault impact factor Quantitative Analysis Model of regression diagnostics in step 4, by it
Logistic regression equation quantitative analyses obtain high-voltage switch gear fault impact factor.
2. the fault impact factor quantitative analysis method according to claim 1 based on high-voltage switch gear, it is characterised in that step
In rapid 2.3, primary data sample is standardized by min-max standardized method, by primary data sample
The codomain of data is mapped in [0,1].
3. the fault impact factor quantitative analysis method according to claim 2 based on high-voltage switch gear, it is characterised in that root
The processing of min-max standardization is carried out to primary data sample according to following formula,
<mrow>
<mi>x</mi>
<mo>*</mo>
<mo>=</mo>
<mfrac>
<mrow>
<mi>x</mi>
<mo>-</mo>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
<mo>-</mo>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, x*The original variable in primary data sample is represented, x represents the variable after standardization, and max represents original change
The upper dividing value in value domain, min represent the floor value of original variable codomain.
4. the fault impact factor quantitative analysis method according to claim 1 based on high-voltage switch gear, it is characterised in that step
The step of also including rejecting independent variable in rapid 3；
After high-voltage switch gear fault impact factor Quantitative Analysis Model is established, conspicuousness rejecting is carried out using method of gradual regression, will
The variable for not meeting AIC index criterions is rejected from regression equation, then it is a pair event to ultimately remain in the independent variable in regression equation
Barrier occurs with the independent variable significantly affected.
5. the fault impact factor quantitative analysis method according to claim 1 based on high-voltage switch gear, it is characterised in that step
In rapid 3, tested by test data set, preferably obtain the tool of optimal high-voltage switch gear fault impact factor Quantitative Analysis Model
Body step is as follows,
A. operational parameter data collection is randomly divided into size identical n parts, n is the positive integer more than 1, is selected in each run
A copy of it is as inspection set, and remaining is all training set, and carries out model construction by Logistic regression algorithms；
B. by step a Repeated ms time, n is the positive integer more than 1, as a result will obtain n m times grader, and pass through each instruction
Practice and examine, the classification accuracy of each grader can be obtained；
C. in calculation procedure b m times of n grader average classification accuracy, computational methods are to take n m times grader standard
The average of true rate；Classification accuracy is found afterwards closest to the object classifiers of average classification accuracy；The object classifiers are just
It is the high-voltage switch gear fault impact factor Quantitative Analysis Model that model preferably obtains.
6. the fault impact factor quantitative analysis method according to claim 1 based on high-voltage switch gear, it is characterised in that step
In rapid 4, regression diagnostics is carried out to the Logistic regression equations of optimal high-voltage switch gear fault impact factor Quantitative Analysis Model, its
The index method of middle diagnosis is likelihood ratio test method.
7. the fault impact factor quantitative analysis method according to claim 1 based on high-voltage switch gear, it is characterised in that step
In rapid 5, the Logistic regression equations of described high-voltage switch gear fault impact factor Quantitative Analysis Model are：
G (x)=ω0+ω1x1+ω2x2+ω3x3 (1-2)
<mrow>
<mi>p</mi>
<mrow>
<mo>(</mo>
<mi>y</mi>
<mo>=</mo>
<mn>1</mn>
<mo>|</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mn>1</mn>
<mo>+</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mi>g</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
</msup>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, x=(x1,x2,x3) represent that independent variable is the vector that each operational parameter data is formed, and p (y=1 | x) represent failure hair
Raw conditional probability；ω0Represent the intercept of regression equation, ω1、ω2、ω3Independent variable x in regression equation is represented respectively1、x2、x3
Coefficient；
It is as follows so as to obtain the result of quantitative analysis,
Failure is made a difference by independent variable coefficient magnitude relation three variables of judgement, and degree is descending to put in order；
Respective fault impact factor is corresponded to according to each operational factor, using independent variable coefficient as weights, passes through weights size
Quantization obtain the size for the influence degree that failure occurs each fault impact factor.
8. the fault impact factor quantitative analysis method according to claim 7 based on high-voltage switch gear, it is characterised in that institute
In the step 5 stated, when obtaining the size for the influence degree that failure occurs each fault impact factor by the quantization of weights size,
When independent variable coefficient is bigger, the influence degree that failure occurs is bigger.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710751628.9A CN107633349A (en) | 2017-08-28 | 2017-08-28 | Fault impact factor quantitative analysis method based on high-voltage switch gear |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710751628.9A CN107633349A (en) | 2017-08-28 | 2017-08-28 | Fault impact factor quantitative analysis method based on high-voltage switch gear |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107633349A true CN107633349A (en) | 2018-01-26 |
Family
ID=61099699
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710751628.9A Pending CN107633349A (en) | 2017-08-28 | 2017-08-28 | Fault impact factor quantitative analysis method based on high-voltage switch gear |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107633349A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109359271A (en) * | 2018-12-21 | 2019-02-19 | 浙江大学 | A kind of deformation of transformer winding degree online test method that logic-based returns |
| CN110287456A (en) * | 2019-06-30 | 2019-09-27 | 张家港宏昌钢板有限公司 | Large coil rolling surface defect analysis method based on data mining |
| CN111190412A (en) * | 2020-01-06 | 2020-05-22 | 珠海格力电器股份有限公司 | Fault analysis method and device, storage medium and terminal |
| CN111611545A (en) * | 2020-05-18 | 2020-09-01 | 国网江苏省电力有限公司电力科学研究院 | Cable aging state evaluation method and device based on principal component analysis and logistic regression |
| CN111783824A (en) * | 2020-05-25 | 2020-10-16 | 北京三清互联科技有限公司 | Method and device for analyzing equipment operation related data |
| CN112365009A (en) * | 2020-10-28 | 2021-02-12 | 国网山东省电力公司电力科学研究院 | Secondary equipment abnormity diagnosis method based on deep learning network |
| CN112557891A (en) * | 2020-11-24 | 2021-03-26 | 广东电网有限责任公司电力科学研究院 | Fault detection method and device for high-voltage switch equipment |
| CN114968990A (en) * | 2022-04-12 | 2022-08-30 | 青岛沃柏斯智能实验科技有限公司 | Design method of diagnosis model for influencing factors of experimental data |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103245911A (en) * | 2013-05-03 | 2013-08-14 | 云南电力试验研究院（集团）有限公司电力研究院 | Breaker fault diagnosis method based on Bayesian network |
| KR101327370B1 (en) * | 2013-06-12 | 2013-11-08 | 김남철 | Protecting cover with drawing type supporter for mobile phone |
| CN104091066A (en) * | 2014-07-04 | 2014-10-08 | 山东大学 | Condition evaluation method for high-voltage circuit breaker |
| CN104316875A (en) * | 2014-11-18 | 2015-01-28 | 广东电网有限责任公司佛山供电局 | On-line monitoring method and system of state of high-voltage circuit breaker |
| CN104484828A (en) * | 2014-11-14 | 2015-04-01 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | High-voltage electrical appliance operation quality evaluation method and system based on logistic regression |
| CN105868770A (en) * | 2016-03-23 | 2016-08-17 | 国网山东省电力公司电力科学研究院 | High-voltage circuit breaker fault diagnosis method based on unsupervised learning model |
| CN106556775A (en) * | 2016-09-27 | 2017-04-05 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | The ultrahigh voltage AC and DC transmission line lightning stroke tripping operation automatic identifying method returned based on Logistic |
-
2017
- 2017-08-28 CN CN201710751628.9A patent/CN107633349A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103245911A (en) * | 2013-05-03 | 2013-08-14 | 云南电力试验研究院（集团）有限公司电力研究院 | Breaker fault diagnosis method based on Bayesian network |
| KR101327370B1 (en) * | 2013-06-12 | 2013-11-08 | 김남철 | Protecting cover with drawing type supporter for mobile phone |
| CN104091066A (en) * | 2014-07-04 | 2014-10-08 | 山东大学 | Condition evaluation method for high-voltage circuit breaker |
| CN104484828A (en) * | 2014-11-14 | 2015-04-01 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | High-voltage electrical appliance operation quality evaluation method and system based on logistic regression |
| CN104316875A (en) * | 2014-11-18 | 2015-01-28 | 广东电网有限责任公司佛山供电局 | On-line monitoring method and system of state of high-voltage circuit breaker |
| CN105868770A (en) * | 2016-03-23 | 2016-08-17 | 国网山东省电力公司电力科学研究院 | High-voltage circuit breaker fault diagnosis method based on unsupervised learning model |
| CN106556775A (en) * | 2016-09-27 | 2017-04-05 | 中国南方电网有限责任公司超高压输电公司检修试验中心 | The ultrahigh voltage AC and DC transmission line lightning stroke tripping operation automatic identifying method returned based on Logistic |
Non-Patent Citations (2)
| Title |
|---|
| 王济川等: "《Logistic回归模型—方法与应用》", 30 September 2001, 高等教育出版社 * |
| 郝爽: ""基于支持向量机的高压断路器机械状态预测算法研究"", 《高压电器》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109359271A (en) * | 2018-12-21 | 2019-02-19 | 浙江大学 | A kind of deformation of transformer winding degree online test method that logic-based returns |
| CN110287456A (en) * | 2019-06-30 | 2019-09-27 | 张家港宏昌钢板有限公司 | Large coil rolling surface defect analysis method based on data mining |
| CN111190412A (en) * | 2020-01-06 | 2020-05-22 | 珠海格力电器股份有限公司 | Fault analysis method and device, storage medium and terminal |
| CN111190412B (en) * | 2020-01-06 | 2021-02-26 | 珠海格力电器股份有限公司 | Fault analysis method and device, storage medium and terminal |
| CN111611545A (en) * | 2020-05-18 | 2020-09-01 | 国网江苏省电力有限公司电力科学研究院 | Cable aging state evaluation method and device based on principal component analysis and logistic regression |
| CN111783824A (en) * | 2020-05-25 | 2020-10-16 | 北京三清互联科技有限公司 | Method and device for analyzing equipment operation related data |
| CN112365009A (en) * | 2020-10-28 | 2021-02-12 | 国网山东省电力公司电力科学研究院 | Secondary equipment abnormity diagnosis method based on deep learning network |
| CN112557891A (en) * | 2020-11-24 | 2021-03-26 | 广东电网有限责任公司电力科学研究院 | Fault detection method and device for high-voltage switch equipment |
| CN112557891B (en) * | 2020-11-24 | 2022-08-02 | 广东电网有限责任公司电力科学研究院 | Fault detection method and device for high-voltage switch equipment |
| CN114968990A (en) * | 2022-04-12 | 2022-08-30 | 青岛沃柏斯智能实验科技有限公司 | Design method of diagnosis model for influencing factors of experimental data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107633349A (en) | Fault impact factor quantitative analysis method based on high-voltage switch gear | |
| CN106019084B (en) | Medium-voltage power distribution network disconnection fault diagnosis method based on power distribution and utilization data correlation | |
| CN112215722B (en) | Dominant instability mode discrimination model construction method and dominant instability mode discrimination method | |
| CN106932670B (en) | A kind of distribution power automation terminal method for diagnosing status based on D-S evidence theory | |
| CN103926490B (en) | A kind of power transformer error comprehensive diagnosis method with self-learning function | |
| CN102289590A (en) | Method for estimating operating state of SF6 high-voltage circuit breaker and intelligent system | |
| CN108073551A (en) | A kind of high-tension switch cabinet on-line fault diagnosis method based on Multi-Agent model | |
| CN110298012B (en) | Test data and operation and maintenance information-based medium and low voltage XLPE cable service life assessment method | |
| CN115549094A (en) | Early warning evaluation method and system for substation equipment of smart power grid | |
| CN107219446A (en) | A kind of gas insulated combined electrical equipment insulation fault recognition methods based on optoacoustic spectroscopy information | |
| CN104091066A (en) | Condition evaluation method for high-voltage circuit breaker | |
| CN110826228B (en) | Regional power grid operation quality limit evaluation method | |
| CN117991708B (en) | Industrial automatic early warning system based on Internet of things | |
| CN107328560A (en) | A kind of load ratio bridging switch diagnostic method and device | |
| CN110580492A (en) | Track circuit fault precursor discovery method based on small fluctuation detection | |
| CN112202597A (en) | Method for evaluating importance of communication network node in low-voltage distribution area | |
| CN104991549A (en) | Track circuit red-light strip default diagnosis method based on FTA and multilevel fuzzy-neural sub-networks | |
| CN109002762A (en) | Method and system for identifying fault current waveform of switching-off and switching-on coil | |
| CN110543658A (en) | Power plant equipment diagnosis method based on big data | |
| CN117708669A (en) | Multi-source data-based adaptive probability neural network high-voltage cable state evaluation method | |
| CN117560300B (en) | Intelligent internet of things flow prediction and optimization system | |
| CN115345392A (en) | Abnormal situation studying and judging and key index screening method for power distribution network | |
| CN111861141B (en) | Power distribution network reliability assessment method based on fuzzy fault rate prediction | |
| CN113505980A (en) | Reliability evaluation method, device and system for intelligent traffic management system | |
| CN109784777B (en) | Power grid equipment state evaluation method based on time sequence information fragment cloud similarity measurement |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210812 Address after: 710077 No. 18 West Second Ring Road, Shaanxi, Xi'an Applicant after: XI'AN HIGH VOLTAGE APPARATUS RESEARCH INSTITUTE Co.,Ltd. Address before: 710075 Shaanxi city of Xi'an province high tech Zone No. 7 Tang Xing Lu Applicant before: CHINA XD ELECTRIC Co.,Ltd. Applicant before: XI'AN HIGH VOLTAGE APPARATUS RESEARCH INSTITUTE Co.,Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180126 |