CN106600452A - Time analysis matrix and clustering analysis-based power distribution network traveling wave fault location method - Google Patents
Time analysis matrix and clustering analysis-based power distribution network traveling wave fault location method Download PDFInfo
- Publication number
- CN106600452A CN106600452A CN201610995316.8A CN201610995316A CN106600452A CN 106600452 A CN106600452 A CN 106600452A CN 201610995316 A CN201610995316 A CN 201610995316A CN 106600452 A CN106600452 A CN 106600452A
- Authority
- CN
- China
- Prior art keywords
- node
- matrix
- row
- time series
- trunk
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 112
- 239000011159 matrix material Substances 0.000 title claims abstract description 89
- 238000000034 method Methods 0.000 title claims abstract description 25
- 230000002159 abnormal effect Effects 0.000 claims abstract description 23
- 230000001052 transient effect Effects 0.000 claims description 17
- 239000000284 extract Substances 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 10
- 230000009466 transformation Effects 0.000 claims description 10
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000005259 measurement Methods 0.000 claims description 3
- 230000000644 propagated effect Effects 0.000 claims description 3
- 241001269238 Data Species 0.000 claims description 2
- 230000003247 decreasing effect Effects 0.000 abstract 1
- 238000007781 pre-processing Methods 0.000 abstract 1
- 230000005540 biological transmission Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000009434 installation Methods 0.000 description 3
- 230000007935 neutral effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 238000007630 basic procedure Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
- G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
-
- 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
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Public Health (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Water Supply & Treatment (AREA)
- Theoretical Computer Science (AREA)
- Locating Faults (AREA)
Abstract
The present invention belongs to the power system automation technical field and relates to a time analysis matrix and clustering analysis-based power distribution network traveling wave fault location method. The method includes the following steps of: the first step, inputting data and performing data preprocessing; the second step, building a time analysis matrix; the third step, performing clustering analysis-based fault section location; the fourth step, checking whether discontinuous nodes and a (k+1)-th column vector satisfy a similarity condition, and identifying abnormal data; the fifth step, deleting abnormal node data; and the sixth step, performing fault point precision location, and outputting a ranging result. According to the time analysis matrix and clustering analysis-based power distribution network traveling wave fault location method of the present invention, an initial wave head time point along a power distribution network line is calculated through utilizing actual terminal node data on the power distribution network line, and a time analysis matrix is constructed; fault section location and abnormal data identification are completed based on the time analysis matrix; and terminal node data most adjacent to a fault point are selected based on the fault section location and abnormal data identification, so that the precise location of the fault point can be completed; and therefore, influence on fault location caused by line length error and wave velocity can be decreased, and the accuracy and reliability of traveling wave location of a power distribution network can be improved.
Description
Technical field
The present invention relates to power system automation technology field, it is a kind of matching somebody with somebody based on time series analyses matrix and cluster analyses
Electrical network traveling wave fault positioning method.
Background technology
China's low and medium voltage distribution network is 3KV to 66KV, more using the operation of isolated neutral or Jing grounding through arc
Mode, is referred to as small current neutral grounding system.Small current neutral grounding system has higher power supply reliability, but distribution line occurs ground connection
After failure, the difficulty of the faint also causing trouble positioning of fault current is larger.Distribution network failure positioning can functionally be divided into two classes：
Fault section location and trouble point are accurately positioned.Fault section location is mainly used in the judgement of fault branch, distribution network failure point
It is accurately positioned the positioning for being mainly used in position of failure point.
At present, it is traveling wave fault location device based on the transmission open acess device of traveling wave principle, at home
It is applied widely on voltage levels circuit, achieves good effect.In recent years, with sampling termination cost drop
Low, public data network transmittability is improved, and multiple research units have carried out power distribution network Travelling Wave Fault Location research both at home and abroad, existing
Power distribution network traveling wave fault positioning method is substantially based on both-end traveling wave positioning, if but only at circuit two ends, installation just cannot be right
Branch trouble is monitored, therefore can install multiple fault location termination i.e. node, the basic procedure of positioning on the line
Two classes can be divided into：(1) reference measure node, reselection are successively selected according to the amplitude size or failure initial time of transient state travelling wave
Two side gusset of reference measure point carries out both-end positioning.(2) primary fault positioning is completed according to circuit two ends outermost node data,
Reselection carries out second positioning apart from the nearest node data in initial alignment trouble point.
But there is problems with Practical Project：
(1) relative to installed in traveling wave fault location device in transformer station, local mounted distribution network failure termination
Limited by installation cost, installation and maintenance condition, operational reliability is relatively low, GPS loses star, AD samplings to be occurred situations such as abnormal
Probability is higher.So as to cause the initial wave head moment that termination is detected with actual value deviation more than 3us (typical traveling wave faults
Positioning time resolving accuracy is 3us), terminal abnormal data are produced, both-end traveling wave positioning needs at least two node datas, arbitrary
Individual node data is all difficult to complete fault location extremely.
(2) as distribution network line traveling wave positioning is generally based on voltage traveling wave, relative to the magnitude of current, voltage traveling wave is being adopted
Affected larger by clutter during sample, while voltage traveling wave causes amplitude distortion and waveform in branch lines total reflection feature
Distortion is more serious so that go wrong by amplitude or the single conditional judgment reference measure point of failure initial time.
Therefore, from practical, through engineering approaches angle, distribution network failure positioning should possess terminal abnormal data identification work(
Can, and comprehensive multipoint data analysis rather than single both-end positioning.
The content of the invention
The invention provides a kind of power distribution network traveling wave fault positioning method based on time series analyses matrix and cluster analyses, gram
Taken the deficiency of above-mentioned prior art, its can in the existing distribution network failure travelling wave positioning method of effectively solving terminal data it is abnormal and
Line length error, velocity of wave produce the problem for affecting to fault location.
The technical scheme is that by following measures to realize：Based on matching somebody with somebody for time series analyses matrix and cluster analyses
Electrical network traveling wave fault positioning method, comprises the following steps：
The first step：Input data, carries out data prediction, and data prediction includes：1) phase-model transformation, using triumphant human relations Bel
Become three-phase voltage of changing commanders and be transformed to Aerial mode component and zero _exit, extract Aerial mode component and be analyzed；2) when lookup failure is initial
Carve, transient signal catastrophe point is detected using wavelet transformation, reach the initial time of each measuring node as transient state travelling wave；3) carry out
The reckoning of circuit trunk node time, draws the initial wave head moment of trunk node, afterwards into second step；
Second step：Time series analyses matrix is built, with each trunk node as traveling wave starting point, segmentation calculates each node along the line
At the initial wave head moment, n × n ranks may make up as matrix element according to the node initial wave head moment that n trunk node is calculated
Time series analyses matrix, time series analyses matrix are shown below：
The 3rd step is entered afterwards；
3rd step：Fault section is positioned based on cluster analyses principle, based on cluster analyses principle to time series analyses
Classification of Matrix, according to the 1st row and the n-th row sample size and failure judgement point position, flow process is as follows：
1) time series analyses matrix the 1st is arranged and the n-th row carries out cluster analyses, the row of time series analyses matrix the 1st and the n-th column vector
Data are divided into two kinds of situations, and the first kind is similar element, and this dvielement is numerically close, and continuous arrangement, and Equations of The Second Kind is gradual change
Data, numerical value difference are larger, extract the close data of numerical value and sort out, are shown below：
2) based on time series analyses matrix samples standard value and quantity, statistical sample quantity and S are positioned to fault section,
Process is as follows：
A, the cluster sample t for taking the row of time series analyses matrix the 1st(k+1)1To tn1, take cluster sample t(k+1)1To tn1In the range of
Any one value is standard value, if cluster sample t(k+1)1To tn1Level off to standard value, sample size S1=n-k；Take the time point
The analysis row cluster sample t of matrix n-th1nTo tkn, take cluster sample t1nTo tknIn the range of any one value be standard value, if cluster
Sample t1nTo tknLevel off to tkn, sample size S2=k；1st row and the n-th row sample size and S=S1+S2=n, then trouble point master
Between dry contact k, k+1, afterwards into the 6th step；
B, the cluster sample t for taking the row of time series analyses matrix the 1stk1To tn1, take cluster sample tk1To tn1In the range of it is any one
Individual value is standard value, if cluster sample tk1To tn1Level off to standard value, sample size S1=n-k+1；Take time series analyses matrix
N row cluster sample t1nTo tkn, take cluster sample t1nTo tknIn the range of any one value be standard value, if cluster sample t1nExtremely
tknLevel off to tkn, sample size S2=k；1st row and the n-th row sample size and S=S1+S2=n+1, then trouble point is positioned at trunk
With the joint of branch road, afterwards into the 6th step；
If the sample size of the 1st row cluster sample of c, time series analyses matrix and the n-th row cluster sample and S<During n, then enter
Enter the 4th step；
4th step：Check whether discontinuous node and+1 column vector of kth meet similarity Condition, recognize abnormal data, mistake
Journey is as follows：
1) discontinuous node is checked, if similar sample is discontinuous in the 1st row or the n-th row, terminal node has abnormal number
According to into the 5th step, if similar sample is continuous in the 1st row or the n-th row, whether inspection+1 column vector of kth meets similarity bar
Part；
2) check whether+1 column vector of kth meets similarity Condition, in S=n-1, trouble point is located on branch road, if but
+ 1 column vector of kth for meeting time series analyses matrix removes tnkIt is outer to meet condition：t1(k+1)To tn(k+1)Level off to tn(k+1), then trouble point
On branch road K+1, into the 6th step, show that neighbour's terminal node has abnormal data if being unsatisfactory for, into the 5th step；
5th step：The data of suppressing exception back end, into the 3rd step；
6th step：Trouble point is accurately positioned, range measurement output；When between trouble point trunk node k, k+1, using master
Dry k, k+1 node data pairing constitutes both-end traveling wave positioning, as follows：
D=lk- (t 'k+1- t 'k)×v/2 (5)
In above formula, d representing faults point and trunk node k distances, lkFor distance, t' between node k, k+1k+1Save for trunk k+1
The initial wave head moment of point, t'kFor the initial wave head moment of trunk k nodes；
When trouble point is on the branch road k, then physical end node k respectively with trunk node k-1 or trunk node k+1
Pairing completes both-end positioning, as follows：
D=lk+bk- (t 'k+1- tk)×v/2 (6)
In above formula, d representing faults point and physical end node k distances, lkFor basic routing line length, b between node k, k+1kFor
Leg length, t'k+1For the initial wave head moment of trunk k+1 nodes, tkFor the initial wave head moment of terminal node k.
The further optimization and/or improvements to foregoing invention technical scheme are presented herein below：
In the above-mentioned first step, circuit trunk node time calculates that process is：After there is ground connection or short circuit in distribution network line, therefore
The transient state travelling wave that barrier is produced can be propagated in electrical network along circuit, and on distribution network line, each terminal can detect transient state travelling wave,
The initial wave head moment of each physical end measuring node of transient state travelling wave arrival is tn；Based on the physical end node initial wave head moment
tnOn the basic routing line that reckoning is obtained, the initial wave head moment of trunk node is t'n；Each trunk euclidean distance between node pair ln；Each branch line
Road length bn；Circuit velocity of wave is set as v, except t6=t'6Outward, the initial wave head moment of each trunk node be shown below：
t'n=tn- bn/v (1)。
In the above-mentioned first step, each trunk node is set as traveling wave starting point, segmentation calculates the initial wave head of each node along the line
At the moment, set with node 1 as starting point, calculate initial the wave head moment such as following formula of node 2,3：
t12=t '1- l1/v
t13=t'1- (l1+l2)/v
(2)
N × n rank time series analyses may make up as matrix element according to the node initial wave head moment that n terminal node is calculated
Matrix, wherein, the diagonal entry t of matrixnn=t'n, time series analyses matrix is shown below：
The present invention calculates the initial wave head moment along the line using each physical end node data on distribution network line and builds
Time series analyses matrix；Fault section location and disorder data recognition are completed based on time series analyses matrix, first, based on cluster analyses
Principle is analyzed to time series analyses matrix, completes fault section location according to sample size and present position, in fault section
On location base, the nearest terminal data in chosen distance trouble point carries out trouble point and is accurately positioned, reduce line length error,
Impact of the velocity of wave to fault location result；Secondly, time series analyses matrix is analyzed based on cluster analyses principle, even if single
There is abnormal data in terminal node, also only affect single sample in column vector, do not affect overall calculation, and pass through time series analyses square
The feature identification abnormal terminals data such as array vector similarity, seriality, it is to avoid which is to the follow-up impact for calculating.
Description of the drawings
Algorithm flow chart of the accompanying drawing 1 for the embodiment of the present invention 1.
Algorithm principle figure of the accompanying drawing 2 for the embodiment of the present invention 1.
Accompanying drawing 3 is the algorithm principle figure of the embodiment of the present invention 2 and embodiment 3.
Specific embodiment
The present invention is not limited by following embodiments, can technology according to the present invention scheme and practical situation determining specifically
Embodiment.
In the present invention, for the ease of description, the description of the relative position relation of each part is according to Figure of description 1
Butut mode being described, such as：The position relationship of upper and lower, left and right etc. is the Butut direction according to Figure of description 1
Come what is determined.
With reference to embodiment and accompanying drawing, the invention will be further described：
Embodiment 1：As shown in accompanying drawing 1,2, power distribution network traveling wave fault that should be based on time series analyses matrix and cluster analyses is fixed
Position method, comprises the following steps：
The first step：Input data, carries out data prediction, and data prediction includes：1) phase-model transformation, using triumphant human relations Bel
Become three-phase voltage of changing commanders and be transformed to Aerial mode component and zero _exit, extract Aerial mode component and be analyzed；2) when lookup failure is initial
Carve, transient signal catastrophe point is detected using wavelet transformation, reach the initial time of each measuring node as transient state travelling wave；3) carry out
The reckoning of circuit trunk node time, draws the initial wave head moment of trunk node, afterwards into second step；
Second step：Time series analyses matrix is built, with each trunk node as traveling wave starting point, segmentation calculates each node along the line
At the initial wave head moment, n × n ranks may make up as matrix element according to the node initial wave head moment that n trunk node is calculated
Time series analyses matrix, time series analyses matrix are shown below：
The 3rd step is entered afterwards；
3rd step：Fault section is positioned based on cluster analyses principle, based on cluster analyses principle to time series analyses
Classification of Matrix, according to the 1st row and the n-th row sample size and failure judgement point position, flow process is as follows：
1) time series analyses matrix the 1st is arranged and the n-th row carries out cluster analyses, the row of time series analyses matrix the 1st and the n-th column vector
Data are divided into two kinds of situations, and the first kind is similar element, and this dvielement is numerically close, and continuous arrangement, and Equations of The Second Kind is gradual change
Data, numerical value difference are larger, extract the close data of numerical value and sort out, are shown below：
2) based on time series analyses matrix samples standard value and quantity, statistical sample quantity and S are positioned to fault section,
Process is as follows：
A, the cluster sample t for taking the row of time series analyses matrix the 1st(k+1)1To tn1, take cluster sample t(k+1)1To tn1In the range of
Any one value is standard value, if cluster sample t(k+1)1To tn1Level off to standard value, sample size S1=n-k；Take the time point
The analysis row cluster sample t of matrix n-th1nTo tkn, take cluster sample t1nTo tknIn the range of any one value be standard value, if cluster
Sample t1nTo tknLevel off to tkn, sample size S2=k；1st row and the n-th row sample size and S=S1+S2=n, then trouble point master
Between dry contact k, k+1, afterwards into the 6th step；
B, the cluster sample t for taking the row of time series analyses matrix the 1stk1To tn1, take cluster sample tk1To tn1In the range of it is any one
Individual value is standard value, if cluster sample tk1To tn1Level off to standard value, sample size S1=n-k+1；Take time series analyses matrix
N row cluster sample t1nTo tkn, take cluster sample t1nTo tknIn the range of any one value be standard value, if cluster sample t1nExtremely
tknLevel off to tkn, sample size S2=k；1st row and the n-th row sample size and S=S1+S2=n+1, then trouble point is positioned at trunk
With the joint of branch road, afterwards into the 6th step；
If the sample size of the 1st row cluster sample of c, time series analyses matrix and the n-th row cluster sample and S<During n, then enter
Enter the 4th step；
4th step：Check whether discontinuous node and+1 column vector of kth meet similarity Condition, recognize abnormal data, mistake
Journey is as follows：
1) discontinuous node is checked, if similar sample is discontinuous in the 1st row or the n-th row, terminal node has abnormal number
According to into the 5th step, if similar sample is continuous in the 1st row or the n-th row, whether inspection+1 column vector of kth meets similarity bar
Part；
2) check whether+1 column vector of kth meets similarity Condition, in S=n-1, trouble point is located on branch road, if but
+ 1 column vector of kth for meeting time series analyses matrix removes tnkIt is outer to meet condition：t1(k+1)To tn(k+1)Level off to tn(k+1), then trouble point
On branch road K+1, into the 6th step, show that neighbour's terminal node has abnormal data if being unsatisfactory for, into the 5th step；
5th step：The data of suppressing exception back end, into the 3rd step；
6th step：Trouble point is accurately positioned, range measurement output；When between trouble point trunk node k, k+1, using master
Dry k, k+1 node data pairing constitutes both-end traveling wave positioning, as follows：
D=lk- (t 'k+1- t 'k)×v/2 (5)
In above formula, d representing faults point and trunk node k distances, lkFor distance, t' between node k, k+1k+1Save for trunk k+1
The initial wave head moment of point, t'kFor the initial wave head moment of trunk k nodes；
When trouble point is on the branch road k, then physical end node k respectively with trunk node k-1 or trunk node k+1
Pairing completes both-end positioning, as follows：
D=lk+bk- (t 'k+1- tk)×v/2 (6)
In above formula, d representing faults point and physical end node k distances, lkFor basic routing line length, b between node k, k+1kFor
Leg length, t'k+1For the initial wave head moment of trunk k+1 nodes, tkFor the initial wave head moment of terminal node k.
Here, the close data of numerical value are extracted in the 3rd step refer to the data for difference being extracted for 3us.
According to actual needs, to the above-mentioned power distribution network traveling wave fault positioning method based on time series analyses matrix and cluster analyses
Make further optimization and/or improvements：
As shown in accompanying drawing 1,2, in the first step, circuit trunk node time calculates that process is：There is ground connection in distribution network line
Or after short circuit, the transient state travelling wave that failure is produced can be propagated in electrical network along circuit, and on distribution network line, each terminal can be detected
To transient state travelling wave, the initial wave head moment of each physical end measuring node of transient state travelling wave arrival is tn；Based on physical end node
Initial wave head moment tnOn the basic routing line that reckoning is obtained, the initial wave head moment of trunk node is t'n；Each trunk euclidean distance between node pair
ln；Each branched line length bn；Circuit velocity of wave is set as v, except t6=t'6Outward, initial the wave head moment such as following formula of each trunk node
It is shown：
t'n=tn- bn/v (1)。
As shown in accompanying drawing 1,2, in the first step, each trunk node is set as traveling wave starting point, segmentation calculates each node along the line
The initial wave head moment, set with node 1 as starting point, calculate node 2,3 initial wave head moment such as following formula：
t12=t '1- l1/v
t13=t'1- (l1+l2)/v
(2)
N × n rank time series analyses may make up as matrix element according to the node initial wave head moment that n terminal node is calculated
Matrix, wherein, the diagonal entry t of matrixnn=t'n, time series analyses matrix is shown below：
The present invention calculates the initial wave head moment along the line using each physical end node data on distribution network line and builds
Time series analyses matrix；Fault section location and disorder data recognition are completed based on time series analyses matrix, first, based on cluster analyses
Principle is analyzed to time series analyses matrix, completes fault section location according to sample size and present position, is positioned in section
On the basis of, the nearest terminal data in chosen distance trouble point carries out trouble point and is accurately positioned, and reduces line length error, velocity of wave
Impact to fault location result；Secondly, time series analyses matrix is analyzed based on cluster analyses principle, even if single terminal
There is abnormal data in node, also only affect single sample in column vector, do not affect overall calculation, and pass through time series analyses rectangular array
The feature identification abnormal terminals data such as vector similitude, seriality, it is to avoid which is to the follow-up impact for calculating.
Embodiment 2：As shown in accompanying drawing 1,2,3, line is led to greatly with Huainan, Anhui 10kV and phantom has been built as template, wherein
The long 19.5km of circuit trunk, each branched line length is as shown above.It is mounted with to determine in line end and 5 main branch road ends
Position termination, sample rate is 1.25MHz with reference to actual device, and model of power transmission system built with reference to actual track structure, employing
Double back line structure, emulation include two kinds of failures, and step is as follows：
On backbone, apart from the about 2.3km of node 3, fault location is calculated according to following steps for trouble point：
The first step：Input data, carries out data prediction, three-phase voltage is carried out phase-model transformation be converted to Aerial mode component and
Zero _exit, extracts Aerial mode component and is analyzed, extract the primary wave that transient state travelling wave reaches each terminal node by wavelet transformation
The head moment, i.e.,：[t1、t2、t3、t4、t5、t6]；According at the beginning of the initial wave head moment of physical end node calculates each node on trunk
At wave head moment beginning, the initial wave head moment of each trunk node can be obtained by formula 1：[t'1、t'2、t'3、t'4、t'5、t'6]=
[12056.8,12043.7,12032.7,12029.8,12049.6,12059.3], simplify three digits after calculating takes, then [t'1、
t'2、t'3、t'4、t'5、t'6]=[56.8,43.6,32.6,29.8,49.6,59.2], afterwards into second step；
Second step：Time series analyses matrix is built, with each trunk node as starting point, combined circuit length segmentation calculates along the line
At the initial wave head moment of each node, time series analyses matrix is constituted, afterwards into the 3rd step, time series analyses matrix is shown below：
3rd step：Fault section is positioned based on cluster analyses principle, based on numerical similarity principle, to the time point
The each column element of analysis matrix carries out cluster analyses, extracts the close data of numerical value and sorts out, as shown in Equation 8, the 1st column matrix element t31Extremely
t61Level off to standard value t31, sample size is 3, and sample position in a matrix is continuous；6th row cluster sample t16Extremely
t36Standard value is close to t16, sample size is 3, then the 1st row and the 6th row sample size sum are：S=6=n, then time series analyses square
Battle array characteristic understands that trouble point is located on basic routing line between node 3,4, afterwards into the 5th step.Here the close data of numerical value are
Data of the difference less than 3us
4th step：Identification abnormal data：Assume that 5 data of node have the error of 5us, then the following institute of time series analyses matrix
Show, the abnormal terminals data of node 5 cause the similar sample of the 1st column vector discontinuous, can identify exception by this feature
Terminal data；
Assume that 4 data of node have the error of 5us, then time series analyses matrix is as follows, and 4 data of node cause S=5=
N-1, and the 4th column data is unsatisfactory for similarity Condition, therefore, 4 terminal data exception of decision node；
5th step：Trouble point is accurately positioned, due to fault section location result show trouble point be located at basic routing line 3,4 it
Between, therefore, using trunk node data t'3、t'4Substitute into both-end computing formula and complete final calculating, with actual fault point position phase
Away from about 90m, it is shown below：
D=(l3- (t '4- t '3) × v)/2=2.21km (11).
Embodiment 3：As shown in accompanying drawing 1,2,3, line is led to greatly with Huainan, Anhui 10kV and phantom has been built as template, wherein
The long 19.5km of circuit trunk, each branched line length is as shown above.It is mounted with to determine in line end and 5 main branch road ends
Position termination, sample rate is 1.25MHz with reference to actual device, and model of power transmission system built with reference to actual track structure, employing
Double back line structure, emulation include two kinds of failures, and step is as follows：
Trouble point is on branch line at the about 1.3km of node 4；
The first step：Input data, carries out data prediction, three-phase voltage is carried out phase-model transformation be converted to Aerial mode component and
Zero _exit, extracts Aerial mode component and is analyzed, extract the primary wave that transient state travelling wave reaches each terminal node by wavelet transformation
The head moment, i.e.,：[t1、t2、t3、t4、t5、t6]；According at the beginning of the initial wave head moment of physical end node calculates each node on trunk
At wave head moment beginning, the initial wave head moment of each trunk node can be obtained：[t'1、t'2、t'3、t'4、t'5、t'6]=[87.4,74.6,
63.6th, 41.4,70.8,81.8], afterwards into second step；
Second step：Time series analyses matrix is built, with each trunk node as starting point, combined circuit length segmentation calculates along the line
At the initial wave head moment of each node, time series analyses matrix is constituted, afterwards into the 3rd step, time series analyses matrix is shown below：
3rd step：Fault section is positioned based on cluster analyses principle, based on numerical similarity principle, to the time point
The each column element of analysis matrix carries out cluster analyses, extracts the close data of numerical value and sorts out, as shown in Equation 13, the 1st column matrix element t51
To t61Level off to standard value t51, sample size is 2；6th row cluster sample t16To t36Standard value is close to t36, sample size is 3,
Then the 1st row and the 6th row sample size sum are：S=6=n-1, and the 4th column vector data meet similarity Condition, by when
Between analysis matrix feature understand that trouble point is located on branched line 4, afterwards into the 4th step, the close data of numerical value are for poor here
Data of the value less than 3us；
4th step：Trouble point is accurately positioned, when trouble point is located on branch road, using physical end data t4With trunk node
t'3Substitute into both-end computing formula and complete final calculating, final calculation result is 115m with actual error, is shown below：
D=(l3+b4- (t4- t '3) × v)/2=1.415km (14).
Above technical characteristic constitutes embodiments of the invention, and which has stronger adaptability and implementation result, can basis
The non-essential technical characteristic of increase and decrease is actually needed, the demand of different situations is met.
Claims (3)
1. a kind of power distribution network traveling wave fault positioning method based on time series analyses matrix and cluster analyses, it is characterised in that：Including
Following steps：
The first step：Input data, carries out data prediction, and data prediction includes：1) phase-model transformation, is converted using triumphant human relations Bel
Three-phase voltage is transformed to into Aerial mode component and zero _exit, Aerial mode component is extracted and is analyzed；2) failure initial time, profit are searched
Transient signal catastrophe point is detected with wavelet transformation, the initial time of each measuring node is reached as transient state travelling wave；3) enter row line
The reckoning of trunk node time, draws the initial wave head moment of trunk node, afterwards into second step；
Second step：Time series analyses matrix is built, with each trunk node as traveling wave starting point, segmentation calculates the initial of each node along the line
Wave head moment, the node initial wave head moment calculated according to n trunk node may make up the time of n × n ranks as matrix element
Analysis matrix, time series analyses matrix are shown below：
The 3rd step is entered afterwards；
3rd step：Fault section is positioned based on cluster analyses principle, based on cluster analyses principle to time series analyses matrix
Classified, according to the 1st row and the n-th row sample size and failure judgement point position, flow process is as follows：
1) time series analyses matrix the 1st is arranged and the n-th row carries out cluster analyses, the row of time series analyses matrix the 1st and the n-th column vector data
It is divided into two kinds of situations, the first kind is similar element, and this dvielement is numerically close, and continuous arrangement, and Equations of The Second Kind is gradient data,
Numerical value difference is larger, extracts the close data of numerical value and sorts out, is shown below：
2) based on time series analyses matrix samples standard value and quantity, statistical sample quantity and S are positioned to fault section, process
It is as follows：
A, the cluster sample t for taking the row of time series analyses matrix the 1st(k+1)1To tn1, take cluster sample t(k+1)1To tn1In the range of it is any
One value is standard value, if cluster sample t(k+1)1To tn1Level off to standard value, sample size S1=n-k；Take time series analyses square
Battle array the n-th row cluster sample t1nTo tkn, take cluster sample t1nTo tknIn the range of any one value be standard value, if cluster sample
t1nTo tknLevel off to tkn, sample size S2=k；1st row and the n-th row sample size and S=S1+S2=n, then trouble point trunk section
Point is between k, k+1, afterwards into the 6th step；
B, the cluster sample t for taking the row of time series analyses matrix the 1stk1To tn1, take cluster sample tk1To tn1In the range of any one value
For standard value, if cluster sample tk1To tn1Level off to standard value, sample size S1=n-k+1；Take time series analyses matrix n-th to arrange
Cluster sample t1nTo tkn, take cluster sample t1nTo tknIn the range of any one value be standard value, if cluster sample t1nTo tkn
Level off to tkn, sample size S2=k；1st row and the n-th row sample size and S=S1+S2=n+1, then trouble point be located at trunk with
The joint of branch road, afterwards into the 6th step；
If the sample size of the 1st row cluster sample of c, time series analyses matrix and the n-th row cluster sample and S<During n, then into
Four steps；
4th step：Check whether discontinuous node and+1 column vector of kth meet similarity Condition, recognize abnormal data, process is such as
Under：
1) discontinuous node is checked, if similar sample is discontinuous in the 1st row or the n-th row, terminal node has abnormal data, enters
Enter the 5th step, if similar sample is continuous in the 1st row or the n-th row, check whether+1 column vector of kth meets similarity Condition；
2) check whether+1 column vector of kth meets similarity Condition, in S=n-1, trouble point is located on branch road, if but meeting
+ 1 column vector of kth of time series analyses matrix removes tnkIt is outer to meet condition：t1(k+1)To tn(k+1)Level off to tn(k+1), then trouble point be located at
On branch road K+1, into the 6th step, show that neighbour's terminal node has abnormal data if being unsatisfactory for, into the 5th step；
5th step：The data of suppressing exception back end, into the 3rd step；
6th step：Trouble point is accurately positioned, range measurement output；When between trouble point trunk node k, k+1, using trunk k,
The pairing of k+1 node datas constitutes both-end traveling wave positioning, as follows：
D=lk-(t'k+1-t'k)×v/2 (5)
In above formula, d representing faults point and trunk node k distances, lkFor distance, t' between node k, k+1k+1For trunk k+1 nodes
Initial wave head moment, t'kFor the initial wave head moment of trunk k nodes；
When trouble point is on branch road k, then physical end node k is matched with trunk node k-1 or trunk node k+1 respectively
Both-end positioning is completed, it is as follows：
D=lk+bk- (t 'k+1- tk)×v/2 (6)
In above formula, d representing faults point and physical end node k distances, lkFor basic routing line length, b between node k, k+1kFor branch road
Length, t'k+1For the initial wave head moment of trunk k+1 nodes, tkFor the initial wave head moment of terminal node k.
2. the power distribution network traveling wave fault positioning method based on time series analyses matrix and cluster analyses according to claim 1,
It is characterized in that in the first step, circuit trunk node time calculates that process is：After there is ground connection or short circuit in distribution network line, failure
The transient state travelling wave of generation can be propagated in electrical network along circuit, and on distribution network line, each terminal can detect transient state travelling wave, temporarily
The initial wave head moment of each physical end measuring node of state traveling wave arrival is tn；Based on the initial wave head moment t of physical end noden
On the basic routing line that reckoning is obtained, the initial wave head moment of trunk node is t'n；Each trunk euclidean distance between node pair ln；Each branched line
Length bn；Circuit velocity of wave is set as v, except t6=t'6Outward, the initial wave head moment of each trunk node be shown below：
t'n=tn- bn/v (1)。
3. the power distribution network Travelling Wave Fault Location side based on time series analyses matrix and cluster analyses according to claim 1 and 2
Method, it is characterised in that in the first step, sets each trunk node as traveling wave starting point, and segmentation calculates the initial wave head of each node along the line
At the moment, set with node 1 as starting point, calculate initial the wave head moment such as following formula of node 2,3：
t12=t '1- l1/v
t13=t'1- (l1+l2)/v (2)
N × n rank time series analyses squares may make up as matrix element according to the node initial wave head moment that n terminal node is calculated
Battle array, wherein, the diagonal entry t of matrixnn=t'n, time series analyses matrix is shown below：
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610995316.8A CN106600452B (en) | 2016-11-11 | 2016-11-11 | Power distribution network traveling wave fault positioning method based on time analysis matrix and cluster analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610995316.8A CN106600452B (en) | 2016-11-11 | 2016-11-11 | Power distribution network traveling wave fault positioning method based on time analysis matrix and cluster analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106600452A true CN106600452A (en) | 2017-04-26 |
| CN106600452B CN106600452B (en) | 2020-07-07 |
Family
ID=58589934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610995316.8A Active CN106600452B (en) | 2016-11-11 | 2016-11-11 | Power distribution network traveling wave fault positioning method based on time analysis matrix and cluster analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106600452B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109061382A (en) * | 2018-08-06 | 2018-12-21 | 长沙理工大学 | A kind of electrical power distribution network fault location method based on the multiterminal traveling wave time difference |
| CN110161376A (en) * | 2019-06-24 | 2019-08-23 | 四川电安智能科技有限公司 | A kind of traveling wave fault moment extraction algorithm |
| CN111610408A (en) * | 2020-05-28 | 2020-09-01 | 广东电网有限责任公司 | Traveling wave fault positioning method, device, equipment and storage medium |
| CN113933649A (en) * | 2021-10-12 | 2022-01-14 | 国网湖南省电力有限公司 | Power distribution network multi-terminal traveling wave fault positioning method utilizing path search |
| CN114325245A (en) * | 2022-01-14 | 2022-04-12 | 湖南大学 | Transmission line fault line selection and positioning method based on traveling wave data deep learning |
| CN116520073A (en) * | 2023-03-17 | 2023-08-01 | 上海交通大学 | Fault positioning method for power supply system of submarine observation network |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020053912A1 (en) * | 2000-07-11 | 2002-05-09 | Murari Saha | Method and device of fault location for distribution networks |
| CN102788926A (en) * | 2012-07-04 | 2012-11-21 | 河南理工大学 | Single-phase ground fault section positioning method of small-current ground system |
| CN105044540A (en) * | 2015-06-30 | 2015-11-11 | 昆明理工大学 | Traveling wave directional protection method for making principal component cluster analysis based on instantaneous power in double-circuit transmission line |
-
2016
- 2016-11-11 CN CN201610995316.8A patent/CN106600452B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020053912A1 (en) * | 2000-07-11 | 2002-05-09 | Murari Saha | Method and device of fault location for distribution networks |
| CN102788926A (en) * | 2012-07-04 | 2012-11-21 | 河南理工大学 | Single-phase ground fault section positioning method of small-current ground system |
| CN105044540A (en) * | 2015-06-30 | 2015-11-11 | 昆明理工大学 | Traveling wave directional protection method for making principal component cluster analysis based on instantaneous power in double-circuit transmission line |
Non-Patent Citations (3)
| Title |
|---|
| 曾祥君等: "基于网络的故障行波定位算法", 《中国电机工程学报》 * |
| 梁睿等: "基于广域行波初始波头时差关系矩阵的配电网故障选线及测距", 《高电压技术》 * |
| 王阳等: "基于图论的配电网故障行波定位新算法", 《电力系统自动化》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109061382A (en) * | 2018-08-06 | 2018-12-21 | 长沙理工大学 | A kind of electrical power distribution network fault location method based on the multiterminal traveling wave time difference |
| CN110161376A (en) * | 2019-06-24 | 2019-08-23 | 四川电安智能科技有限公司 | A kind of traveling wave fault moment extraction algorithm |
| CN110161376B (en) * | 2019-06-24 | 2021-05-28 | 四川电安智能科技有限公司 | Traveling wave fault time extraction algorithm |
| CN111610408A (en) * | 2020-05-28 | 2020-09-01 | 广东电网有限责任公司 | Traveling wave fault positioning method, device, equipment and storage medium |
| CN113933649A (en) * | 2021-10-12 | 2022-01-14 | 国网湖南省电力有限公司 | Power distribution network multi-terminal traveling wave fault positioning method utilizing path search |
| CN113933649B (en) * | 2021-10-12 | 2023-09-12 | 国网湖南省电力有限公司 | Multi-terminal traveling wave fault positioning method for power distribution network by utilizing path searching |
| CN114325245A (en) * | 2022-01-14 | 2022-04-12 | 湖南大学 | Transmission line fault line selection and positioning method based on traveling wave data deep learning |
| CN116520073A (en) * | 2023-03-17 | 2023-08-01 | 上海交通大学 | Fault positioning method for power supply system of submarine observation network |
| CN116520073B (en) * | 2023-03-17 | 2024-01-30 | 上海交通大学 | Fault positioning method for power supply system of submarine observation network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106600452B (en) | 2020-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106600452A (en) | Time analysis matrix and clustering analysis-based power distribution network traveling wave fault location method | |
| CN104297638B (en) | Power distribution grid fault positioning method with high precision and low cost | |
| CN104330708B (en) | Fault location method for wide area traveling wave signal below power grid | |
| CN104898021A (en) | Fault line selecting method based on k-means cluster analysis for power distribution network | |
| CN102830328A (en) | Distributed fault location method for T-circuit | |
| CN105425107A (en) | Method and system for active power distribution network fault diagnosis and location | |
| CN102331548B (en) | Single-phase earth fault positioning method for distribution network cable | |
| CN102967779B (en) | Identifying method of distribution parameters of transmission line | |
| CN107632236A (en) | A kind of single outgoing-feeder line one-end fault ranging method based on the identification of opposite end bus back wave | |
| CN117192292B (en) | Lightning grounding electrode line fault distance measurement method and system | |
| CN108693446B (en) | A kind of Fault Locating Method of non-synchronous sampling power grid transient state travelling wave modulus time difference | |
| CN110133444A (en) | A kind of Fault Locating Method based on positive sequence voltage variable quantity, apparatus and system | |
| CN111766470B (en) | Fault positioning method and system for high-voltage direct-current transmission line and direct-current transmission line | |
| CN105699849A (en) | Voltage sag estimation method based on quantum-behaved particle swarm optimization algorithm | |
| CN114113890A (en) | Multi-terminal fusion power distribution network fault positioning method based on traveling wave modulus transmission time difference | |
| Wei et al. | Traveling-wave-based fault location algorithm for star-connected hybrid multi-terminal HVDC system | |
| CN115097253A (en) | MMC-HVDC direct current transmission line fault distance measurement method and system | |
| CN109521330A (en) | A kind of transmission line malfunction travelling wave ranging method based on the prediction of ARIMA wave head | |
| CN113884818B (en) | Method for accurately estimating arrival time of fault traveling wave of power distribution network based on LSTM | |
| CN103941149A (en) | Fault location method for electrified railway traction network | |
| CN102183709A (en) | Method of determining fault point of power grid and severity of fault | |
| CN107179473A (en) | A kind of power transmission line fault locating method | |
| CN117421976A (en) | Wind power plant current collection line fault positioning monitoring method and system based on ensemble learning | |
| CN103972889A (en) | Distribution line impedance online identification method | |
| CN111537832A (en) | Line fault positioning method, terminal and system for multi-terminal flexible direct current transmission system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |