CN109409607A - The method of power circuit tripping rate with lightning strike prediction based on composite factor - Google Patents
The method of power circuit tripping rate with lightning strike prediction based on composite factor Download PDFInfo
- Publication number
- CN109409607A CN109409607A CN201811290424.0A CN201811290424A CN109409607A CN 109409607 A CN109409607 A CN 109409607A CN 201811290424 A CN201811290424 A CN 201811290424A CN 109409607 A CN109409607 A CN 109409607A
- Authority
- CN
- China
- Prior art keywords
- lightning strike
- attribute
- tripping rate
- value
- rate
- 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
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The method for the power circuit tripping rate with lightning strike prediction based on composite factor that the invention discloses a kind of, it is by studying influence of the different factors to tripping rate with lightning strike, using rough set theory, attribute reduction is carried out to Multiple factors using tripping rate with lightning strike as decision attribute, to effectively extract several factors vital to tripping rate with lightning strike, and it is extracted the correlation rule of one group of prediction tripping rate with lightning strike on this basis, to realize the quick predict of tripping rate with lightning strike.Prediction technique of the present invention is simple, it is only necessary to measure the correlation in relation to factor, as condition, be matched by one group of correlation rule to tripping rate with lightning strike, can predict tripping rate with lightning strike, avoid various cumbersome calculating process.Prediction technique in the present invention also has generalization and portability, for the other accident in electric system, can also be predicted according to its related factor.
Description
Technical field
The present invention relates to technical field of power systems, especially a kind of power circuit tripping rate with lightning strike based on composite factor
The method of prediction.
Background technique
The reliability of transmission line of electricity is most important to the safe and stable operation of entire electric system, therefore, once power line
Trip accident occurs for road, will bring serious harm to the safe and stable operation of entire electric system, and according to statistics, when lightning stroke draws
The main reason for sending out high-voltage fence tripping, the number that trip accident occurs because of lightning stroke for annual high-voltage electric power circuit accounts about total jump
The 50%-60% of lock number, since topography is more complicated, the number to trip is higher in some places, and therefore, research lightning stroke is jumped
Lock rate is very necessary to power system power supply reliability.
Currently, being concentrated mainly on the following aspects to the research of tripping rate with lightning strike: first: analysis tripping rate with lightning strike generates
The reason of.These analyses only are being possible to make explanations to the factor that trip-out rate has an impact, and can not really predict thunder
It hits in the range and many factors of trip-out rate, which influences to be necessary, which is negligible.Second: lightning stroke trip
The calculation method and modification method of rate.For a long time, many people attempt to find fixed a formula or set pattern to calculate
Tripping rate with lightning strike out, however because the variation and each department condition difference of each factor value, cause numerical value that can change, are
, also there are various modification methods, but correct anyway in this, and tripping rate with lightning strike is since influence factor is numerous, handle of having no idea
All factors are considered.
By finding to the summary and recent research studied in the past: tripping rate with lightning strike is related to several factors, is based on
Analysis and observation is carried out to these factor values, can probably obtain the range of tripping rate with lightning strike.Based on this discovery, the present invention is proposed
A method of the power circuit tripping rate with lightning strike prediction based on composite factor.
Summary of the invention
A kind of power circuit lightning stroke based on composite factor is provided it is an object of the invention to solve above-mentioned technical problem
The method of trip-out rate prediction, can be realized the quick predict of tripping rate with lightning strike.
In order to solve the above-mentioned technical problem, the present invention adopts the following technical scheme:
A method of the power circuit tripping rate with lightning strike prediction based on composite factor, which is characterized in that including following step
It is rapid:
Step 1 obtains the influence factor for influencing tripping rate with lightning strike；Note tripping rate with lightning strike is D, influences the shadow of tripping rate with lightning strike
The factor of sound is followed successively by C1, C2 ..., Ck；
Step 2 is repeatedly tested, and X1, X2, X3 ..., Xn are successively denoted as；And record tripping rate with lightning strike and influence factor
Correlation；
Real number is expressed each influence factor correlation progress classifying and dividing using equiprobability criterion method by step 3
Correlation is converted into qualitative attribute value；
Referred herein to " influence factor " be tripping rate with lightning strike " attribute "；By multiple test, it will obtain continuous multiple
About the value of attribute Ci (1≤i≤k), the value of whole attribute Ci (1≤i≤k) is considered as a sequence, is estimated using cuclear density
Algorithm draws the probability density curve of the sequence, and corresponding probability density function is denoted as fi(x)；Remember CiminAnd CimaxRespectively
For for years test in, the minimum value and maximum value about attribute Ci measured, CiaWith CibBelong to real interval [Cimin,
Cimax] and Cib>Cia, and meetIt is abbreviated as S1=S2=S3, wherein
S1, S2, S3 are respectively fi(x) in section [Cimin, Cia)、[Cia,Cib]、(Cib,Cimax] on integral；It is this to attribute value
Equiprobability classification draws method and is known as " equiprobability criterion "；The Ci determined based on the criterionmin、Cia、Cib、Cimax；
According to the property of probability it is recognised that the value for waiting the attribute Ci of the tripping rate with lightning strike of prediction falls in [Cimin,
Cia)、[Cia,Cib]、(Cib,Cimax] probability in three sections be it is equal, the attribute value about Ci as follows is formulated with this
Classifying and dividing standard:
If tripping rate with lightning strike falls in section (0, Ci about the attribute value of Cia) or [Cia,Cib] or (Cib,+∞) on when, then
Its classification value is 1 or 2 or 3；
Likewise, also carrying out classifying and dividing using attribute value of the equiprobability criterion to tripping rate with lightning strike D；
Step 4, using the influence factor of tripping rate with lightning strike as conditional attribute, using tripping rate with lightning strike as decision attribute, with above-mentioned
Classifying and dividing result be attribute value, establish tripping rate with lightning strike decision table；
Test of many times X1, X2, X3 ..., the influence factor and tripping rate with lightning strike of Xn carries out classification by the method for step 3 and draws
Point, obtain each attribute to deserved attribute value, mutually deserved decision table is shown in Table 1, and in table, tripping rate with lightning strike D is decision attribute, influences thunder
The factor C1, C2 ..., Ck for hitting trip-out rate are conditional attribute, Vij、ViDAttribute value of the respectively Xi about Cj and D, and have Vij、ViD
∈ { 1,2,3 }；I=1,2 ..., n；J=1,2 ..., k；
1 tripping rate with lightning strike decision table of table
| U | C1 | C2 | … | Ck | D |
| x1 | V11 | V12 | … | V1k | V1D |
| x2 | V21 | V22 | … | V2k | V2D |
| … | … | … | … | … | … |
| xn | Vn1 | Vn2 | … | Vnk | VnD |
Table 1 describe decision information system a S=<U, A, V, f>,
Wherein, U={ X1, X2, X3 ..., Xn } is domain, indicates the set of continuous n test；
A is the union of conditional attribute collection C={ C1, C2 ..., Ck } and decision set { D }；
V is domain object X1, X2, X3 ..., the union of the codomain of each attribute of Xn, herein { 1,2,3 } V=；
F is X1, X2, X3 ..., and information MAP function of the Xn about attribute is here the attribute value classifying and dividing in step 3
Method；
Step 5 is directed to decision table, carries out conditional attribute to the Relative Reduced Concept of decision attribute based on rough set theory, obtains
One decision abridged table；
Step 6 obtains one group of correlation rule for predicting tripping rate with lightning strike by decision abridged table；
The decision abridged table obtained according to step 5 describes in words out, obtains one group of association rule of tripping rate with lightning strike
Then, while the confidence level of every correlation rule is calculatedAnd have
In formula, Xf(f=1,2,3 ... p) be the domain U' after reduction in the equivalence class partition set U'/C of conditional attribute C
Element, while the subset of the element or original domain U={ X1, X2, X3 ..., Xn }, p are the element number in U'/C, p≤n
|Xf| it is set XfGesture, i.e. U' is to the element number in the equivalence class partition set of conditional attribute C；
Yt(t=1,2,3 ... q) be the domain U' after attribute reduction in the equivalence class partition set U'/D of conditional attribute D
Element, while the subset of the element or original domain U={ X1, X2, X3X ... Xn }, q is the element number in U'/D, q≤n,
|Yt∩Xf| it is set Yt∩XfGesture, i.e. set Yt∩XfThe number of middle element, Yt∩XfFor XfWith YtIntersection；
The correlation rule form of tripping rate with lightning strike is as follows:
Rule(c1,1)∧(C2,2)∧(C3,2)→(D,2)；If this correlation rule is in artificial intelligence
Production rule expression, then are as follows: if C1=1 and C2=2 and C3=2, D=2；The Rules control is 1；
Step 7 carries out storage and management to the data information of tripping rate with lightning strike and correlation rule in the form of unique file；
Step 8, association rule and relevant knowledge, complete the prediction to tripping rate with lightning strike；
Corresponding correlation rule is recalled from computer, and rule match is carried out to correlation rule group according to test result, and
It is an area in the predicted value of this tripping rate with lightning strike using the conclusion of the correlation rule of successful match as tripping rate with lightning strike predicted value
Between be worth rather than a determining number；
The predicted value of tripping rate with lightning strike is carried out display output by step 9 in the display of computer.
Further, influence factor described in step 1 include the ground resistance of shaft tower, the shielding angle of lightning conducter, tower height,
Insulator the piece number, ground elevation.
Further, test is repeatedly carried out in step 2 to refer to using electric power special equipment to the shadow for influencing tripping rate with lightning strike
Ring factor tested and acquired, wherein be acquired data when repeatedly should measure and be averaged.
Further, conditional attribute C={ C1, C2 ..., Ck } is carried out to decision attribute D according to rough set theory in step 5
Relative Reduced Concept, the specific method is as follows:
The positive domain POS of C of D is found out respectivelyC(D) and C-CjPositive domainAnd judge POSC(D) withIt is
It is no equal, the C if equaljFor in C relative to D can reduction attribute, i.e. CjBe it is unnecessary in C, otherwise claim CjIt is opposite in C
In D can not reduction attribute, i.e. CjIt is necessary in C.Update attribute collection and decision table, i.e. identical sample merge, then after
It is continuous to calculate POSC(D) andUntil can be deleted without attribute again, a Relative Reduced Concept of conditional attribute collection C is obtained,
It is denoted as redD(C), the element in the set contributes to the important indicator of prediction tripping rate with lightning strike, redD(C) belong to influence thunder in
The element for hitting trip-out rate factor is known as " great influence parameter ".
Further, data information described in step 7 and correlation rule include the tripping rate with lightning strike that is obtained in step 2 with
And " the great influence parameter " for being considered predicting tripping rate with lightning strike for being respectively worth and being denoted as Le, obtain in step 5 of influence factor
redD(C), what is obtained in step 6 is used to predict the correlation rule group of tripping rate with lightning strike；Data information and correlation rule are concentrated
It is stored in some specified physical space of computer, the computer physical space that name stores this file is knowledge base, according to
The frequency lambda for the correct conclusion that every correlation rule provides in practical applications, dynamic adjust the confidence level of each rule in knowledge base,
Method be byAs the confidence level of respective rule, whereinFor the confidence level for obtaining rule in step 6.
The beneficial effects of the present invention are:
1, the present invention is by studying influence of the different factors to tripping rate with lightning strike, using rough set theory, with lightning stroke trip
Rate be decision attribute to Multiple factors carry out attribute reduction, thus effectively extract it is vital to tripping rate with lightning strike it is several because
Element, and it is extracted the correlation rule of one group of prediction tripping rate with lightning strike on this basis, to realize the quick pre- of tripping rate with lightning strike
It surveys.
2, prediction technique of the present invention is simple, it is only necessary to the correlation in relation to factor is measured, as condition, by lightning stroke
One group of correlation rule of trip-out rate is matched, and can predict tripping rate with lightning strike, avoids various cumbersome calculating process.
3, expense of the present invention is low, it is only necessary to measure to the influence factor of needs, save the cost.
4, the prediction technique in the present invention has generalization and portability, for the other accident in electric system,
It can be predicted according to its related factor.Method in the present invention can extend use.
Detailed description of the invention
Fig. 1 is the classifying and dividing schematic diagram of the influence factor of the tripping rate with lightning strike based on equiprobability criterion；
Fig. 2 is the classifying and dividing schematic diagram of the tripping rate with lightning strike based on equiprobability criterion.
Specific embodiment
The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:
The method of power circuit tripping rate with lightning strike prediction based on composite factor of the invention, comprising the following steps:
Step 1: obtaining the parameter (influence factor) for influencing tripping rate with lightning strike.
The parameter for influencing lightning stroke trip has very much, such as: the ground resistance of shaft tower, the shielding angle of lightning conducter, tower height, insulation
Sub-pieces number, ground elevation etc., each parameter are referred to as an attribute of tripping rate with lightning strike, comprehensively choose the above ginseng as far as possible
Number.Tripping rate with lightning strike refers to that in the case where thunderstorm days Td=40, tripping is secondary because of caused by lightning stroke every year for the route of 100km
Number, unit are " secondary/(100km.40 thunderstorm day) ", and note tripping rate with lightning strike is D, influence the influence factor of tripping rate with lightning strike successively
For C1, C2 ..., Ck；
Step 2: repeatedly being tested, record the correlation of its tripping rate with lightning strike and influence factor.
It is repeatedly tested, is successively denoted as X1, X2, X3 ..., Xn；Record the correlation of tripping rate with lightning strike and influence factor.
Wherein, the correlation of influence factor should be measured by power engineering special equipment, in measurement, should repeatedly be measured and be taken
Average value.Alternatively, to Electric Design department obtain, Electric Design department carry out overhead line structures, power circuit design when there are
Specific data, it is negotiable to learn specific data.
Step 3: using equiprobability criterion method, each influence factor correlation is subjected to classifying and dividing, by real number expression
Correlation is converted into qualitative classification value.
Herein, " influence factor " is referred to as " attribute " of tripping rate with lightning strike.By multiple test, it will obtain continuous more
The value of whole attribute Ci (1≤i≤k) is considered as a sequence, using cuclear density by a value about attribute Ci (1≤i≤k)
Estimation algorithm draws the probability density curve (cuclear density estimation algorithm is well known method in mathematical statistics) of the sequence, and will be corresponding
Probability density function is denoted as fi(x)；Remember CiminAnd CimaxRespectively for years test in, measure about attribute Ci most
Small value and maximum value, CiaWith CibBelong to real interval [Cimin, Cimax] and Cib>Cia, and meetIt is abbreviated as S1=S2=S3, wherein S1, S2, S3 are respectively fi(x) exist
Section [Cimin, Cia)、[Cia,Cib]、(Cib,Cimax] on integral；This equiprobability classification to attribute value is drawn method and is known as
" equiprobability criterion "；The Ci determined based on the criterionmin、Cia、Cib、Cimax, as shown in Figure 1.
According to the property of probability it is recognised that the value for waiting the attribute Ci of the tripping rate with lightning strike of prediction falls in [Cimin,
Cia)、[Cia,Cib]、(Cib,Cimax] probability in three sections be it is equal, the attribute value about Ci as follows is formulated with this
Classifying and dividing standard:
If tripping rate with lightning strike falls in section (0, Ci about the attribute value of Cia) or [Cia,Cib] or (Cib,+∞) on when, then
Its classification value is 1 or 2 or 3；
The principle of above-mentioned division methods is: by repeatedly measuring to certain attribute of tripping rate with lightning strike, it will be appreciated that
The entire probability distribution situation of its attribute value, then using the attribute value of random value equiprobably fall in three sections as target into
Row divides, so that classification value is converted by the attribute value that real number is expressed, this also means that random any one year tripping rate with lightning strike
It is equiprobable, this classifying and dividing method to attribute value referred to as " equiprobability standard that classification value about the attribute, which takes 1 or 2 or 3,
Then ".
Likewise, also carrying out classifying and dividing using attribute value of the equiprobability criterion to tripping rate with lightning strike D.By repeatedly trying
It tests, multiple tripping rate with lightning strike can be obtained, the value of this multiple tripping rate with lightning strike is considered as a sequence, is drawn using cuclear density estimation algorithm
The density curve of the sequence is made, and corresponding probability density function is denoted as f1 (T), and obtains three equivalent area S1=S2
=S3, as shown in Figure 2.In figure, Tmin、TmaxThe respectively minimum value and maximum value of tripping rate with lightning strike, S1, S2, S3 be respectively by
F1 (T) and section [Tmin, Ta)、[Ta,Tb]、(Tb,Tmax] area that defines, and have S1=S2=S3.If not considering other spies
Different situation, then tripping rate with lightning strike can equiprobably fall in [Tmin, Ta)、[Ta,Tb]、(Tb,Tmax] within three sections of sections.It makes accordingly
Determine the classifying and dividing standard of D attribute value: when tripping rate with lightning strike belongs to section [Tmin, Ta) or [Ta,Tb] or (Tb,Tmax] when, classification
Value takes 1 or 2 or 3 respectively；Fig. 2 is the tripping rate with lightning strike classifying and dividing schematic diagram based on equiprobability criterion.
Step 4: using the influence factor of tripping rate with lightning strike as conditional attribute, using tripping rate with lightning strike as decision attribute, with above-mentioned
Classifying and dividing result be attribute value, establish tripping rate with lightning strike decision table.
Test of many times X1, X2, X3 ..., the influence factor and tripping rate with lightning strike of Xn carries out classification by the method for step 3 and draws
Point, obtain each attribute to deserved attribute value (classification value), mutually deserved decision table is shown in Table 1, and in table, tripping rate with lightning strike D is decision category
Property, the factor C1 of tripping rate with lightning strike is influenced, C2 ..., Ck are conditional attribute, Vij、ViDAttribute value of the respectively Xi about Cj and D
(classification value), and have Vij、ViD∈ { 1,2,3 }；I=1,2 ..., n；J=1,2 ..., k.
1 tripping rate with lightning strike decision table of table
| U | C1 | C2 | … | Ck | D |
| x1 | V11 | V12 | … | V1k | V1D |
| x2 | V21 | V22 | … | V2k | V2D |
| … | … | … | … | … | … |
| xn | Vn1 | Vn2 | … | Vnk | VnD |
Table 1 describe decision information system a S=<U, A, V, f>,
Wherein, U={ X1, X2, X3 ..., Xn } is domain, indicates the set of continuous n test.
A is the union of conditional attribute collection C={ C1, C2 ..., Ck } and decision set { D }.
V is domain object X1, X2, X3 ..., the union of the codomain of each attribute of Xn, herein { 1,2,3 } V=
F is X1, X2, X3 ..., and information MAP function of the Xn about attribute is here the attribute value classifying and dividing in step 3
Method.
Step 5: being directed to decision table, conditional attribute is carried out to the Relative Reduced Concept (attribute of decision attribute based on rough set theory
Relative Reduced Concept be well known method in classical rough set theory), obtain a decision abridged table.
For step 4 decision table obtained, conditional attribute C={ C1, C2 ..., Ck } (letter is carried out according to rough set theory
Claim C) to the Relative Reduced Concept of decision attribute D (abbreviation D), the specific method is as follows:
The positive domain POS of C of D is found out respectivelyC(D) and C-CjPositive domainAnd judge POSC(D) withIt is
It is no equal, the C if equaljFor in C relative to D can reduction attribute, i.e. CjBe it is unnecessary in C, otherwise claim CjIt is opposite in C
In D can not reduction attribute, i.e. CjIt is necessary in C.Update attribute collection and decision table, i.e. identical sample merge, then after
It is continuous to calculate POSC(D) andUntil can be deleted without attribute again, a Relative Reduced Concept of conditional attribute collection C is obtained,
It is denoted as redD(C), the element in the set contributes to the important indicator of prediction tripping rate with lightning strike, redD(C) belong to influence thunder in
The element for hitting trip-out rate factor is known as " great influence parameter ".
Step 6: one group of correlation rule for predicting tripping rate with lightning strike is obtained by decision abridged table.
The decision abridged table obtained according to step 5 describes in words out, obtains one group of association rule of tripping rate with lightning strike
Then, while according to formula (1) confidence level of every correlation rule is calculatedAnd have
In formula, Xf(f=1,2,3 ... p) be the domain U' after reduction in the equivalence class partition set U'/C of conditional attribute C
Element, while the subset of the element or original domain U={ X1, X2, X3 ..., Xn }, p are the element number in U'/C, p≤n
|Xf| it is set XfGesture, i.e. U' is to the element number in the equivalence class partition set of conditional attribute C.
Yt(t=1,2,3 ... q) be the domain U' after attribute reduction in the equivalence class partition set U'/D of conditional attribute D
Element, while the subset of the element or original domain U={ X1, X2, X3X ... Xn }, q is the element number in U'/D, q≤n,
|Yt∩Xf| it is set Yt∩XfGesture, i.e. set Yt∩XfThe number of middle element, Yt∩XfFor XfWith YtIntersection.
The correlation rule form of tripping rate with lightning strike is as follows:
Rule(c1,1)∧(C2,2)∧(C3,2)→(D,2)；If this correlation rule is in artificial intelligence
Production rule expression, then are as follows: if C1=1 and C2=2 and C3=2, D=2；The Rules control is 1.
Confidence level reflects the credibility of rule, and confidence level, is to measure (to be generated by the condition class of correlation rule in other words
The former piece or condition of formula rule) release Decision Classes (i.e. the consequent or conclusion of production rule) trusted degree, expression
Degree of support of the condition part to conclusion part, the rule that confidence level is 1 is Deterministic rules, the attribute value 1 and 2 in rule
It is all the qualitative description to attribute value, the two respectively corresponds to a real number interval.
Step 7: storage and management being carried out to the dependency rule of tripping rate with lightning strike and relevant knowledge in the form of unique file.
Each value of the tripping rate with lightning strike and influence factor that are obtained in storing step 2 in the form of unique file, is denoted as Le, step
What is obtained in rapid 5 is considered predicting the red of " the great influence parameter " of tripping rate with lightning strikeD(C), being used for for obtaining in step 6 is pre-
Survey the correlation rule group of tripping rate with lightning strike.These information are centrally stored in some specified physical space of computer, are named
The computer physical space for storing this file is knowledge base.The correct conclusion provided in practical applications according to every correlation rule
Frequency lambda, dynamic adjust knowledge base in each rule confidence level, method be byAs the confidence level of respective rule,
WhereinFor the confidence level for obtaining rule in step 6.
Step 8: association rule and relevant knowledge complete the prediction to tripping rate with lightning strike.
Using electric power special equipment, the factor for influencing tripping rate with lightning strike is tested and acquired, data are being acquired
When, it should repeatedly measure, be averaged, finally arrange data.Corresponding correlation rule, root are recalled from computer
Rule match is carried out to correlation rule group according to test result, and using the conclusion of the correlation rule of successful match as tripping rate with lightning strike
Predicted value, here, the predicted value of tripping rate with lightning strike are an interval values rather than a determining number.
Step 9: exporting the predicted value of tripping rate with lightning strike.
The predicted value of tripping rate with lightning strike is subjected to display output in the display of computer.
Embodiment 1
Step 1: choosing sample, obtain the parameter for influencing tripping rate with lightning strike.
Choosing the ground resistance of shaft tower, the shielding angle of lightning conducter, tower height, insulator the piece number, ground elevation is to influence lightning stroke
Five factors of trip-out rate.And remember tripping rate with lightning strike, the ground resistance of shaft tower, the shielding angle of lightning conducter, tower height, sub-pieces
Number, ground elevation are followed successively by D, C1, C2, C3, C4, C5.
Step 2: being tested, record the correlation of its tripping rate with lightning strike and influence factor.
It is measured using power engineering special equipment, in measurement, should repeatedly measure and be averaged.Alternatively, to electric power
Design department obtain, Electric Design department carry out overhead line structures, power circuit design when there are specific data, it is negotiable to learn
Specific data.
| Sample 1 | Sample 2 | Sample 3 | Sample 4 | Sample 5 | Sample 6 | Sample 7 | Sample 8 | |
| Ground resistance | 19.5 | 15.3 | 17.4 | 12 | 8.1 | 22.8 | 17.8 | 10.8 |
| Shielding angle | 20.3 | 23.2 | 26.8 | 18.9 | 14.7 | 20.9 | 14.2 | 21.5 |
| Tower height | 36.5 | 33.7 | 25.6 | 39.2 | 37.3 | 42.7 | 47.5 | 29.2 |
| Insulator the piece number | 13 | 14 | 8 | 15 | 14 | 14 | 16 | 11 |
| Ground elevation | 21 | 27 | 13 | 9 | 21 | 20 | 11 | 23 |
| Tripping rate with lightning strike | 0.238 | 0.226 | 0.195 | 0.163 | 0.198 | 0.279 | 0.182 | 0.264 |
Step 3: using equiprobability criterion as the classifying and dividing method of each attribute value, converting the attribute value that real number is expressed to
Classification value.
According to the data of upper table, the corresponding measurement sequence of each attribute is drawn using cuclear density estimation algorithm respectively and is formed by
Then probability density curve is determined two homalographic separations of each curve by equiprobability criterion.It i.e. can by all properties value
Interal separation where energy is as follows: at three subintervals 1,2,3
Step 4: tripping rate with lightning strike decision table
| U | C1 | C2 | C3 | C4 | C5 | D |
| X1 | 2 | 2 | 2 | 2 | 3 | 2 |
| X2 | 2 | 2 | 2 | 2 | 3 | 2 |
| X3 | 2 | 3 | 1 | 2 | 2 | 2 |
| X4 | 1 | 2 | 2 | 3 | 1 | 2 |
| X5 | 1 | 1 | 2 | 2 | 3 | 2 |
| X6 | 2 | 2 | 2 | 2 | 2 | 3 |
| X7 | 2 | 1 | 3 | 3 | 1 | 2 |
| X8 | 1 | 2 | 2 | 2 | 3 | 3 |
Step 5: carrying out attribute reduction with rough set theory.
U={ X1, X2, X3, X4, X5, X6, X7, X8 }
U/ind (C1)={ { X1, X2, X3, X6, X7 }, { X4, X5, X8 } }
U/ind (C2)={ { X1, X2, X4, X6, X8 }, { X3 }, { X5, X7 } }
U/ind (C3)={ { X1, X2, X4, X5, X6, X8 }, { X3 }, { X7 } }
U/ind (C4)={ { X1, X2, X3, X5, X6, X8 }, { X4, X7 } }
U/ind (C5)={ { X1, X2, X5, X8 }, { X3, X6 }, { X4, X7 } }
U/ind (C)={ { X1, X2 }, { X3 }, { X4 }, { X5 }, { X6 }, { X7 }, { X8 } }
U/ind (D)={ { X1, X2, X3, X4, X5, X7 }, { X6, X8 } }
U/ind (C-C1)={ { X1, X2, X8 }, { X3 }, { X4 }, { X5 }, { X6 }, { X7 } }
U/ind (C-C2)={ { X1, X2 }, { X3 }, { X4 }, { X5, X8 }, { X6 }, { X7 } }
U/ind (C-C3)={ { X1, X2 }, { X3 }, { X4 }, { X5 }, { X6 }, { X7 }, { X8 } }
U/ind (C-C4)={ { X1, X2 }, { X3 }, { X4 }, { X5 }, { X6 }, { X7 }, { X8 } }
U/ind (C-C5)={ { X1, X2, X6 }, { X3 }, { X4 }, { X5 }, { X7 }, { X8 } }
U/ind (C-C3-C4)={ { X1, X2 }, { X3 }, { X4 }, { X5 }, { X6 }, { X7 }, { X8 } }
POSC(D)={ X1, X2 } ∪ { X3 } ∪ { X4 } ∪ { X5 } ∪ { X6 } ∪ { X7 } ∪ { X8 }
={ X1, X2, X3, X4, X5, X6, X7, X8 }
Dependency degree k=γ between property set D and CC(D)=| POSC(D) |/| U |=1O declared attribute collection C=C1, C2,
C3, C4, C5 }={ ground resistance, the shielding angle of lightning conducter, tower height, insulator the piece number, ground elevation } divide tripping rate with lightning strike
Class is sufficiently to gather, and it is adequately that C is relative to D that product sample { X1, X2, X3, X4, X5, X6, X7, X8 }, which is classified as D, according to C
It is able to carry out reduction.
As known from the above,C1 can not independent reduction relative to D
C2 can not independent reduction relative to D
C3 can independent reduction relative to D
C4 can independent reduction relative to D
C5 can not independent reduction relative to D
C3 and C4 can reduction simultaneously relative to D
In summary, C1, C2, C5 not can be carried out reduction relative to D, so " important indicator " that influences tripping rate with lightning strike is
Ground resistance, the shielding angle of lightning conducter, ground elevation.
Step 6: one group of correlation rule for predicting tripping rate with lightning strike is obtained by decision abridged table
| U | C1 | C2 | C5 | D |
| X1 | 2 | 2 | 3 | 2 |
| X2 | 2 | 2 | 3 | 2 |
| X3 | 2 | 3 | 2 | 2 |
| X4 | 1 | 2 | 1 | 2 |
| X5 | 1 | 1 | 3 | 2 |
| X6 | 2 | 2 | 2 | 3 |
| X7 | 2 | 1 | 1 | 2 |
| X8 | 1 | 2 | 3 | 3 |
The confidence level that correlation rule is calculated using correlation formula, obtains following correlation rule:
Rule1:(C1,2)Λ(C2,2)Λ(C5,3)→(D,2)；Confidence level is 1
Rule2:(C1,2)Λ(C2,3)Λ(C5,2)→(D,2)；Confidence level is 1
Rule3:(C1,1)Λ(C2,2)Λ(C5,1)→(D,2)；Confidence level is 1
Rule4:(C1,1)Λ(C2,1)Λ(C5,3)→(D,2)；Confidence level is 1
Rule5:(C1,2)Λ(C2,2)Λ(C5,2)→(D,3)；Confidence level is 1
Rule6:(C1,2)Λ(C2,1)Λ(C5,1)→(D,2)；Confidence level is 1
Rule7:(C1,1)Λ(C2,2)Λ(C5,3)→(D,3)；Confidence level is 1
By above-mentioned 7 rules, describe in words as follows:
Rule1: if pole tower ground resistance, in [15,24] section, and shielding angle is in [15,25] section, and
Ground elevation [20 ,+∞) in section, then tripping rate with lightning strike is in [0.158,0.263] section.The confidence level of the rule is
1。
Rule2: if pole tower ground resistance in [15,24] section, and shielding angle [25 ,+∞) in section, and
Ground elevation is in [12,20] section, then tripping rate with lightning strike is in [0.158,0.263] section.The confidence level of the rule is
1。
Rule3: if pole tower ground resistance (0,15] in section, and shielding angle is in [15,25] section, and ground
Face inclination angle (0,12] in section, then tripping rate with lightning strike is in [0.158,0.263] section.The confidence level of the rule is 1.
Rule4: if pole tower ground resistance (0,15] in section, and shielding angle (0,15] in section, and ground
Face inclination angle [20 ,+∞) in section, then tripping rate with lightning strike is in [0.158,0.263] section.The confidence level of the rule is 1.
Rule5: if pole tower ground resistance, in [15,24] section, and shielding angle is in [15,25] section, and
Ground elevation in [12,20] section, then tripping rate with lightning strike [0.263 ,+∞) in section.The confidence level of the rule is 1.
Rule6: if pole tower ground resistance in [15,24] section, and shielding angle (0,15] in section, and ground
Face inclination angle (0,12] in section, then tripping rate with lightning strike is in [0.158,0.263] section.The confidence level of the rule is 1.
Rule7: if pole tower ground resistance (0,15] in section, and shielding angle is in [15,25] section, and ground
Face inclination angle [20 ,+∞) in section, then tripping rate with lightning strike [0.263 ,+∞) in section.The confidence level of the rule is 1.
Step 7: storage and management being carried out to the dependency rule of tripping rate with lightning strike and relevant knowledge in the form of unique file.
Each value of the tripping rate with lightning strike and influence factor that are obtained in storing step 2 in the form of unique file obtains in step 5
What is obtained is considered predicting " important indicator " of tripping rate with lightning strike: ground resistance, the shielding angle of lightning conducter, ground elevation., step 6
7 correlation rules for being used to predict tripping rate with lightning strike of middle acquisition.
Step 8: association rule and relevant knowledge complete the prediction to tripping rate with lightning strike.
Using electric power special equipment, the factor for influencing tripping rate with lightning strike is tested and acquired, data are being acquired
When, it should mostly this measure, be averaged, finally obtaining data is 16.7 ohm of ground resistance, the shielding angle 19.9 of lightning conducter,
Ground elevation 20, extracts correlation rule in knowledge base, obtain prediction result [0.263 ,+∞).
Step 9: exporting the predicted value of tripping rate with lightning strike.
The predicted value of tripping rate with lightning strike is subjected to display output in the display of computer.
In conclusion the contents of the present invention are not limited in the above embodiments, those skilled in the art can be
It is proposed other embodiments within technological guidance's thought of the invention, but these embodiments be included in the scope of the present invention it
It is interior.
Claims (5)
1. a kind of method of the power circuit tripping rate with lightning strike prediction based on composite factor, which comprises the following steps:
Step 1 obtains the influence factor for influencing tripping rate with lightning strike；Note tripping rate with lightning strike is D, influence the influence of tripping rate with lightning strike because
Element is followed successively by C1, C2 ..., Ck；
Step 2 is repeatedly tested, and X1, X2, X3 ..., Xn are successively denoted as；And record the phase of tripping rate with lightning strike and influence factor
Pass value；
Step 3, using equiprobability criterion method, each influence factor correlation is subjected to classifying and dividing, the correlation that real number is expressed
Value is converted into qualitative attribute value；
Referred herein to " influence factor " be tripping rate with lightning strike " attribute "；By multiple test, it will obtain it is continuous it is multiple about
The value of whole attribute Ci (1≤i≤k) is considered as a sequence, using cuclear density estimation algorithm by the value of attribute Ci (1≤i≤k)
The probability density curve of the sequence is drawn, and corresponding probability density function is denoted as fi(x)；Remember CiminAnd CimaxRespectively exist
For years in test, the minimum value and maximum value about attribute Ci measured, CiaWith CibBelong to real interval [Cimin, Cimax]
And Cib>Cia, and meetBe abbreviated as S1=S2=S3, wherein S1, S2,
S3 is respectively fi(x) in section [Cimin, Cia)、[Cia,Cib]、(Cib,Cimax] on integral；This equiprobability to attribute value
Classification draws method and is known as " equiprobability criterion "；The Ci determined based on the criterionmin、Cia、Cib、Cimax；
According to the property of probability it is recognised that the value for waiting the attribute Ci of the tripping rate with lightning strike of prediction falls in [Cimin, Cia)、
[Cia,Cib]、(Cib,Cimax] probability in three sections be it is equal, classified as follows about the attribute value of Ci with this to formulate
The criteria for classifying:
If tripping rate with lightning strike falls in section (0, Ci about the attribute value of Cia) or [Cia,Cib] or (Cib,+∞) on when, then its point
Class value is 1 or 2 or 3；
Likewise, also carrying out classifying and dividing using attribute value of the equiprobability criterion to tripping rate with lightning strike D；
Step 4, using the influence factor of tripping rate with lightning strike as conditional attribute, using tripping rate with lightning strike as decision attribute, with above-mentioned point
Class division result is attribute value, establishes tripping rate with lightning strike decision table；
Test of many times X1, X2, X3 ..., the influence factor and tripping rate with lightning strike of Xn carries out classifying and dividing by the method for step 3,
Each attribute is obtained to deserved attribute value, mutually deserved decision table is shown in Table 1, and in table, tripping rate with lightning strike D is decision attribute, influences to be struck by lightning
The factor C1, C2 ..., Ck of trip-out rate are conditional attribute, Vij、ViDAttribute value of the respectively Xi about Cj and D, and have Vij、ViD∈
{ 1,2,3 }；I=1,2 ..., n；J=1,2 ..., k；
1 tripping rate with lightning strike decision table of table
Table 1 describe decision information system a S=<U, A, V, f>,
Wherein, U={ X1, X2, X3 ..., Xn } is domain, indicates the set of continuous n test；
A is the union of conditional attribute collection C={ C1, C2 ..., Ck } and decision set { D }；
V is domain object X1, X2, X3 ..., the union of the codomain of each attribute of Xn, herein { 1,2,3 } V=；
F is X1, X2, X3 ..., and information MAP function of the Xn about attribute is here the attribute value classifying and dividing side in step 3
Method；
Step 5 is directed to decision table, carries out conditional attribute to the Relative Reduced Concept of decision attribute based on rough set theory, obtains one
Decision abridged table；
Step 6 obtains one group of correlation rule for predicting tripping rate with lightning strike by decision abridged table；
The decision abridged table obtained according to step 5 describes in words out, obtains one group of correlation rule of tripping rate with lightning strike, together
When calculate the confidence level of every correlation ruleAnd have
In formula, Xf(f=1,2,3 ... p) be the domain U' after reduction to the member in the equivalence class partition set U'/C of conditional attribute C
Element, while the subset of the element or original domain U={ X1, X2, X3 ..., Xn }, p are the element number in U'/C, p≤n | Xf|
For set XfGesture, i.e. U' is to the element number in the equivalence class partition set of conditional attribute C；
Yt(t=1,2,3 ... q) be the domain U' after attribute reduction to the member in the equivalence class partition set U'/D of conditional attribute D
Element, while the subset of the element or original domain U={ X1, X2, X3X ... Xn }, q are the element number in U'/D, q≤n, | Yt
∩Xf| it is set Yt∩XfGesture, i.e. set Yt∩XfThe number of middle element, Yt∩XfFor XfWith YtIntersection；
The correlation rule form of tripping rate with lightning strike is as follows:
Rule(c1,1)∧(C2,2)∧(C3,2)→(D,2)；If this correlation rule is with the generation in artificial intelligence
Formula regular expression, then are as follows: if C1=1 and C2=2 and C3=2, D=2；The Rules control is 1；
Step 7 carries out storage and management to the data information of tripping rate with lightning strike and correlation rule in the form of unique file；
Step 8, association rule and relevant knowledge, complete the prediction to tripping rate with lightning strike；
Corresponding correlation rule is recalled from computer, and rule match, and general are carried out to correlation rule group according to test result
Conclusion with successful correlation rule is an interval value in the predicted value of this tripping rate with lightning strike as tripping rate with lightning strike predicted value
A rather than determining number；
The predicted value of tripping rate with lightning strike is carried out display output by step 9 in the display of computer.
2. the method for the power circuit tripping rate with lightning strike prediction described in accordance with the claim 1 based on composite factor, feature exist
In, influence factor described in step 1 include the ground resistance of shaft tower, the shielding angle of lightning conducter, tower height, insulator the piece number,
Face inclination angle.
3. the method for the power circuit tripping rate with lightning strike prediction described in accordance with the claim 1 based on composite factor, feature exist
In, in step 2 repeatedly carry out test refer to using electric power special equipment to influence tripping rate with lightning strike influence factor test
And acquisition, wherein be acquired data when repeatedly should measure and be averaged.
4. the method for the power circuit tripping rate with lightning strike prediction described in accordance with the claim 1 based on composite factor, feature exist
According to rough set theory progress conditional attribute C={ C1, C2 ..., Ck } to the Relative Reduced Concept of decision attribute D, tool in step 5
Body method is as follows:
The positive domain POS of C of D is found out respectivelyC(D) and C-CjPositive domainAnd judge POSC(D) withWhether phase
Deng the C if equaljFor in C relative to D can reduction attribute, i.e. CjBe it is unnecessary in C, otherwise claim CjFor in C relative to D not
Can reduction attribute, i.e. CjIt is necessary in C.Update attribute collection and decision table, i.e., identical sample merge, then proceed to calculate
POSC(D) andUntil that can be deleted without attribute again, a Relative Reduced Concept of conditional attribute collection C is obtained, is denoted as
redD(C), the element in the set contributes to the important indicator of prediction tripping rate with lightning strike, redD(C) belong to influence lightning stroke in jump
The element of lock rate factor is known as " great influence parameter ".
5. the method for the power circuit tripping rate with lightning strike prediction based on composite factor, feature exist according to claim 4
In, data information described in step 7 and correlation rule include each of the tripping rate with lightning strike obtained in step 2 and influence factor
The red for being considered predicting " the great influence parameter " of tripping rate with lightning strike for being worth and being denoted as Le, obtain in step 5D(C), in step 6
What is obtained is used to predict the correlation rule group of tripping rate with lightning strike；Data information and correlation rule are concentrated to certain for being stored in computer
A specified physical space, the computer physical space that name stores this file is knowledge base, according to every correlation rule in reality
The frequency lambda of correct conclusion provided in the application of border, dynamic adjust the confidence level of each rule in knowledge base, method be by
As the confidence level of respective rule, whereinFor the confidence level for obtaining rule in step 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811290424.0A CN109409607A (en) | 2018-10-31 | 2018-10-31 | The method of power circuit tripping rate with lightning strike prediction based on composite factor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811290424.0A CN109409607A (en) | 2018-10-31 | 2018-10-31 | The method of power circuit tripping rate with lightning strike prediction based on composite factor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109409607A true CN109409607A (en) | 2019-03-01 |
Family
ID=65470777
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811290424.0A Pending CN109409607A (en) | 2018-10-31 | 2018-10-31 | The method of power circuit tripping rate with lightning strike prediction based on composite factor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109409607A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110398666A (en) * | 2019-08-29 | 2019-11-01 | 南方电网科学研究院有限责任公司 | A kind of Fault Diagnosis Method for Distribution Networks based on relay protection timing information feature |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104239968A (en) * | 2014-09-02 | 2014-12-24 | 浙江大学 | Short-term load predicting method based on quick fuzzy rough set |
| CN104462799A (en) * | 2013-11-27 | 2015-03-25 | 河北工业大学 | Relay individual working life predicting and screening method based on early life performance |
| JP2015057942A (en) * | 2010-04-28 | 2015-03-26 | 株式会社東芝 | Power consumption management system, power consumption management device, power consumption management method, central supply power management device, and supply power management method to be used for the same system |
-
2018
- 2018-10-31 CN CN201811290424.0A patent/CN109409607A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015057942A (en) * | 2010-04-28 | 2015-03-26 | 株式会社東芝 | Power consumption management system, power consumption management device, power consumption management method, central supply power management device, and supply power management method to be used for the same system |
| CN104462799A (en) * | 2013-11-27 | 2015-03-25 | 河北工业大学 | Relay individual working life predicting and screening method based on early life performance |
| CN104239968A (en) * | 2014-09-02 | 2014-12-24 | 浙江大学 | Short-term load predicting method based on quick fuzzy rough set |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110398666A (en) * | 2019-08-29 | 2019-11-01 | 南方电网科学研究院有限责任公司 | A kind of Fault Diagnosis Method for Distribution Networks based on relay protection timing information feature |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108427041B (en) | Lightning early warning method, system, electronic equipment and storage medium | |
| Zhu et al. | Time series shapelet classification based online short-term voltage stability assessment | |
| CN112114579B (en) | Industrial control system safety measurement method based on attack graph | |
| Li et al. | Rainfall and earthquake-induced landslide susceptibility assessment using GIS and Artificial Neural Network | |
| WO2021077729A1 (en) | Lightning prediction method | |
| CN105677791A (en) | Method and system used for analyzing operating data of wind generating set | |
| CN107257351A (en) | One kind is based on grey LOF Traffic anomaly detections system and its detection method | |
| CN109374986A (en) | A kind of Lightning Location Method and system based on clustering and grid search | |
| Aligholian et al. | Event detection in micro-pmu data: A generative adversarial network scoring method | |
| CN109061774A (en) | A kind of thunderstorm core relevance processing method | |
| CN109829627A (en) | A kind of safe confidence appraisal procedure of Electrical Power System Dynamic based on integrated study scheme | |
| CN106651031A (en) | Lightning stroke flashover early warning method and system based on historical information | |
| CN110472775A (en) | A kind of series case suspect's foothold prediction technique | |
| CN109118105A (en) | The risk analysis method and system of power grid mass-sending failure under mountain fire disaster | |
| CN110716998B (en) | Fine scale population data spatialization method | |
| CN109409607A (en) | The method of power circuit tripping rate with lightning strike prediction based on composite factor | |
| Mabayoje et al. | Gain ratio and decision tree classifier for intrusion detection | |
| CN104181205A (en) | Composite material damage identification method and system thereof | |
| CN109193703A (en) | Consider the electric power system transient stability key feature selection method of classification lack of uniformity | |
| Lei et al. | Statistical feature selection of narrowband RCS sequence based on greedy algorithm | |
| CN114707912A (en) | Power grid risk detection method, device and equipment | |
| Zhang et al. | Particle swarm optimization pattern recognition neural network for transmission lines faults classification | |
| CN111160712B (en) | User electricity consumption parameter adjusting method and device | |
| Xiaolan et al. | Determination of the weight values of assessment indexes of website based on AHP-take the website of university library as an example | |
| CN110426612A (en) | A kind of two-stage type transformer oil paper insulation time domain dielectric response characteristic quantity preferred method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190301 |