CN106373032A - Distribution network high-fault region identification method - Google Patents

Abstract

The invention discloses a distribution network high-fault region identification method. The method comprises the following steps: performing fuzzy identification on distribution network fault positions on the basis of multisource data; based on historical load data and meteorological monitoring data, performing distribution network fault power failure load loss evaluation; performing grid division on a geological view and realizing mapping between faults and geological grids; calculating a fault influence statistical index of each geological grid in a given time period and quantifying fault influences in each grid area; according to each fault influence statistical index, performing coloring so as to generate a fault influence color spot graph; and bringing forward corresponding auxiliary decision-making advice to a high-fault region. According to the invention, power failure loss quantification and fuzzy positioning are performed on vast distribution network faults, the mapping is performed with the geological grids, the fault influences of the geological grids are statistically analyzed, and the fault influence color spot graph is formed, such that refined and customized auxiliary decision-making advance is provided for distribution network scheduling, operation and maintenance.

Description

Distribution Network Failure based on big data region occurred frequently discrimination method

Technical field

The present invention relates to a kind of region occurred frequently of the Distribution Network Failure based on big data discrimination method, belong to distribution safety with can By property analysis field.

Background technology

In traditional Distribution Network Failure management work, often studied and judged just for wall scroll fault, and entered on this basis Row region dimension (province, city, county) and time dimension (year, month, day) Distribution Network Failure quantity statistics.These statistical result granularities are relatively Slightly it is impossible to provide effective aid decision to distribution scheduling fortune inspection.

Single electric network fault has certain occasionality, but from the point of view of one section of long period, distribution rack is thin The fault rate in the region that weak or O＆M falls behind is higher than average level, and presents certain regularity.Therefore, by history event Barrier carries out big data analysis, effectively identifies the high risk zone of Frequent Troubles, and takes corresponding technology or management means Reduce the probability that failure risk occurs, be of great significance for improving Reliability of Power Supplying Net Work tool.

The continuous propulsion built with intelligent grid, the level of IT application of power distribution network is constantly lifted, and current Jiangsu Power Grid is Ems (EMS), pms (production management system), dms (distribution management system) and electricity consumption acquisition system have been done step-by-step Comprehensive covering.

Content of the invention

In view of the shortcomings of the prior art, it is an object of the present invention to provide a kind of Distribution Network Failure district occurred frequently based on big data Domain discrimination method, is quantified and fuzzy positioning by the Distribution Network Failure of magnanimity has been carried out with loss of outage, and is carried out with geographic grid Mapping, and then carry out statistical analysiss for the fault impact of geographic grid, forms fault impact mottle figure, thus for distribution tune Degree, O＆M, maintenance provide the aid decision suggestion becoming more meticulous and customizing.

To achieve these goals, the present invention is to realize by the following technical solutions:

Distribution Network Failure based on the big data region occurred frequently discrimination method of the present invention, including following step:

(1) it is directed to every fault together, fuzzy diagnosis is carried out to Distribution Network Failure position based on multi-source data；Based on historical load Data, weather data carry out the assessment of Distribution Network Failure power failure load loss；

(2) stress and strain model is carried out to geographic view, and realize the mapping of fault and geographic grid；

(3) calculate preset time section [t_s,t_e] in each geographic grid fault impact statistical indicator, to each net region Internal fault impact is quantified；

(4) affect value of statistical indicant according to each grid fault geographic grid is coloured, thus generating fault impact color Speckle figure；

(5) region occurred frequently for fault, proposes corresponding aid decision suggestion.

In step (1), the method carrying out fuzzy diagnosis to described Distribution Network Failure position is as follows:

(1a) according to regular expression, the abort situation from electric network fault statistical report system, districts and cities' sole duty made a report on is civilian This information is parsed, and identifies device numbering and device name, and searches corresponding device in production management system pms, from And determine abort situation (pos_x, pos_y), and labelling abort situation type is 0, if None- identified, goes to step (2a)；

(2a) the block switch remote signalling searching faulty line from distribution management system dms divides position to record, if this circuit All block switches are mounted with distributing automation apparatus, find out two block switches of separating brake at first in line fault time range ss₁、ss₂, and labelling abort situation type is 1；If circuit does not install distributing automation apparatus, go to step (3a)；

(3a) search the load data that fault feeder has all distribution transformers under its command from power information acquisition system, according to Electric current falls zero duration, determines the power failure duration of each distribution transforming, and the distribution transforming the longest of power failure duration is added to distribution transforming set pb_ In set, and labelling abort situation type is 2.

In step (2), the method that described geographic view is carried out with stress and strain model is as follows:

Geographic area be based on first quartile coordinate system, the square that geographic area is len with the length of side carry out cutting it is assumed that Amount to m row n row square and cover given geographic area, wherein, m 3, n > 3, it is g (0,0) that the grid in the lower left corner is numbered, Then being located at the grid that xth arranges, y line position is put and numbering is g (x, y).

In step (2), described fault is as follows with the mapping method of geographic grid:

Define event of failure list ft_list (x, y) of geographic grid g (x, y), occur in this grid for storage Event of failure, each record comprises with properties: time of failure occure_time, fault feeder feeder_info, Abort situation gps coordinate x-axis pos_x, abort situation gps coordinate y-axis pos_y, fault outage loss load loss, fault type Type, abort situation fuzziness amb and breakdown loss weighted value weight；

Distribution Network Failure is traveled through, for fault ft_i,

Case1: this location of fault type is 0, if its fault occurs position (pos_x_i,pos_y_i) ∈ g (x, y), then Ft will be recorded_iInsertion ft_list (x, y), wherein, ft_i.Amb=0, ft_i.Weight=1, ft_i.pos_x=pos_x_i, ft_i.pos_y=pos_y_i；

Case2: this location of fault type is 1,

Case2.1: if two block switch ss₁、ss₂And the circuit between them is all located among g (x, y), then will Record ft_iInsertion ft_list (x, y), wherein, ft_i.Amb=0, ft_i.Weight=1；

Case2.2: by two block switch ss₁、ss₂And the circuit between them is referred to as ft_section, if ft_ Section is located at different geographic grid g (x₁,y₁)～g (x_n,y_n) in, fault is inserted into these ground according to different weights In the error listing of reason grid；For geographic grid g (x_j,y_j), ft will be recorded_iInsertion ft_list (x_j,y_j), wherein, ft_i. Amb=1, ft_i.

Case3: this location of fault type is 2,

Case3.1: if pb_set and the circuit between them are all located among g (x, y), then will record ft_iInsertion Ft_list (x, y), wherein, ft_i.Amb=0, ft_i.Weight=1；

Case3.2: by pb_set and the track section that connects these distribution transformings is referred to as ft_section, if ft_ Section is located at different geographic grid g (x₁,y₁)～g (x_n,y_n) in, fault is inserted into these ground according to different weights In the error listing of reason grid；For geographic grid g (x_j,y_j), ft will be recorded_iInsertion ft_list (x_j,y_j), wherein, ft_i. Amb=1, ft_i.

In step (3), described fault impact statistical indicator includes fault frequency, fault outage impact accumulated value, event Barrier has a power failure affects line length average and fault outage impact 10kv capacity of distribution transform average；

For grid g (x, y), the computational methods that described fault affects statistical indicator are as follows:

(1b) fault frequency

$ft\_count={\mathrm{\σ}}_{i=1}^{n^{\′}}{\mathrm{ft}}_i.weight$

Wherein, n ' is to occur Primary fault number of times in grid；

(2b) fault outage impact accumulated value

$ft\_sum=\underset{i=1}{\overset{n^{\′}}{\σ}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)$

(3b) fault outage impact line length average

$ft\_avg\_line=\frac{{\mathrm{\σ}}_{i=1}^{n^{\′}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)}{{\mathrm{\σ}}_{j=1}^{m^{\′}}(length\_grid({\mathrm{line}}_j))}$

Wherein, length_grid (line_j) it is length in grid g (x, y) for the circuit j；

M ' is the sum of distribution transforming comprising in grid or the sum of the feeder line interlocking with grid, line_jFor j-th strip in grid Feeder line；

(4b) fault outage impact 10kv capacity of distribution transform average

$ft\_avg\_tr=\frac{{\mathrm{\σ}}_{i=1}^{n^{\′}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)}{{\mathrm{\σ}}_{j=1}^{m^{\′}}\left(volumn\right({\mathrm{tr}}_j))}$

Wherein, volumn (tr_j) for distribution transformer j capacity, wherein, tr_jFor jth platform distribution transforming in grid.

In step (4), described fault impact mottle map generalization method is as follows:

After selected statistical regions, statistical time range, fault type, statistical indicator, calculate the fault statistics value of each geographic grid Two-dimensional array g_stat [x] [y]；

Statistics obtains 95 probits of this array, 85 probits and 30 probits, respectively as high-risk, serious, common and The threshold value that lower region divides；

Each geographic grid is coloured according to corresponding grade, thus forming fault impact mottle figure.

In step (5), described aid decision suggestion is as follows:

(1c) high risk zone, and based on overhead transmission line, then suggestion carries out insulating, cabling reconstructing or carries out distribution certainly Dynamicization upgrading；

(2c) high risk zone, and based on cable but power distribution automation low degree, then advise carrying out distribution to circuit automatic Change upgrading；

(3c) high risk zone in, and statistical indicator affects line length average for fault outage, then suggestion increases this region Line sectionalizing；

(4c) high risk zone in, and fault type is external force destruction, then the communication with unit in charge of construction is strengthened in suggestion, and increases The dynamics that this region is patrolled and examined；

(5c) high risk zone in, and fault type is ageing equipment, then this area equipment investigation is strengthened in suggestion, upgrades in time Defect or aging equipment；

(6c) high risk zone in, and fault type is tree line contradiction, then advise periodically carrying out this region tree trimming.

The present invention can realize Distribution Network Failure zone location, and the loss of outage that fault is caused carries out accurate evaluation.? On the basis of this, Distribution Network Failure is mapped with geographic grid, realize the breakdown loss value based on geographic grid and add up to quantify, and Render geographic grid using the mode of mottle figure to be pointed out, thus providing upgrading, transformation, strengthening for scheduling, Yun Jiandeng department Spy such as patrols at precision and the aid decision suggestion customizing.

Brief description

Fig. 1 is the region occurred frequently of the Distribution Network Failure based on the big data discrimination method workflow diagram of the present invention；

Fig. 2 is the schematic diagram that geographic grid of the present invention divides with position mark.

Specific embodiment

Technological means, creation characteristic, reached purpose and effect for making the present invention realize are easy to understand, with reference to Specific embodiment, is expanded on further the present invention.

Referring to Fig. 1, the domain occurred frequently of the Distribution Network Failure based on big data discrimination method, Distribution Network Failure is carried out with fuzzy positioning, and The loss of outage that fault is caused carries out accurate evaluation；On this basis, Distribution Network Failure is mapped with geographic grid, realized Breakdown loss value based on geographic grid adds up to quantify, and renders geographic grid using the mode of mottle figure and pointed out, for adjusting Degree, Yun Jiandeng department provide upgrading, transformation, strengthen that spy such as patrols at the precision and the aid decision of customization supports.

In order to effectively use this method, should ensure that possess the geographic view in certain region, power distribution network topological structure, distribution therefore Barrier list of thing, Distribution Network Failure reporting information, feeder line and distribution transforming machine account information, feeder line and distribution transformer load data, weather monitoring number According to etc..

This method includes following five steps:

(1) the intelligent fuzzy identification of Distribution Network Failure position and fault outage loss are accurately assessed.

The intelligent fuzzy identification of Distribution Network Failure position, particularly as follows:

(1a) according to regular expression, the abort situation text made a report on according to districts and cities' sole duty from electric network fault reporting system Information is parsed, and pattern includes: " #xx ring main unit ", " #xx bar ", " #xx on-pole switch ", " #xx cut cable ", " #xx joins Change ", " #xx feeder pillar " etc., identify key message (as device numbering, title etc.), and make a look up corresponding setting in pms Standby, so that it is determined that fault approximate location (pos_x, pos_y), and labelling abort situation type is 0.If None- identified, go to Step (2a)；

(2a) the block switch remote signalling searching faulty line from dms divides position to record, if all block switches of this circuit It is mounted with two distant (or three distant) device it is assumed that line fault time range is [t₁,t₂], find out in this time period separating brake at first Two block switch ss₁、ss₂, and labelling abort situation type is 1.If circuit does not install distributing automation apparatus, go to Step (3a)；

(3a) search the load data that this feeder line has all distribution transformers under its command from power information acquisition system, according to electricity Stream is that zero duration determines its power failure duration, and the distribution transforming the longest of power failure duration is added in distribution transforming set pb_set, and Labelling abort situation type is 2.Judge power failure duration except falling zero according to electric current, stop telegram in reply event also dependent on distribution transforming and come really Determine power failure duration (if possess event acquisition and report mechanism).

Distribution Network Failure power failure load loss assessment particularly as follows:

Using history distribution transformer load, weather history live data, built according to the similar day algorithm based on human comfort Distribution transformer load Short-term Forecasting Model；According to the weather data of fault period of right time, estimate period each distribution transforming loss of being out of order Load curve, and integrate and calculate loss of outage electricity.Each distribution transforming loss load is added up, obtains feeder line power failure load loss Value.Specifically can refer to patent: the Distribution Network Failure loss of outage appraisal procedure of meter and distributed new, Application No.: 201610151200.6.

(2) geographic view carries out stress and strain model and geographic grid and fault correlation mapping.

Geographic view stress and strain model, particularly as follows:

The square that geographic area is len with the length of side carries out cutting, can enter according to actual needs Mobile state adjustment it is proposed that It is worth for 1km～10km.Assume that amounting to m row n row square covers given geographic area (m > 3, n > 3), by the net in the lower left corner It is g (0,0) that lattice are numbered.For geographic area g (x, y), (0, m), (0, n), eight grids of its direct neighbor are permissible for y ∈ for x ∈ Be represented sequentially as g (x-1, y), grid (x-1, y-1), g (x, y-1), g (x-1, y+1), g (x+1, y), grid (x+1, y-1), g (x, y+1), g (x+1, y+1), specifically as shown in Figure 2.Geographic view can be using state's net gis platform or Baidu map, high moral map Open map platform Deng socialization.

Based on the Fault Mapping of geographic grid, particularly as follows:

Define event of failure list ft_list (x, y) of geographic grid g (x, y) first, it stores and occurs in this grid Event of failure, each record comprises with properties:

1) time of failure occure_time；

2) fault feeder feeder_info；

3) abort situation gps coordinate x-axis pos_x

4) abort situation gps coordinate y-axis pos_y

5) fault outage loss load loss；

6) fault type type.Remarks: comprise vile weather, tree line contradiction, ageing equipment, external force destruction, user's reason, The type such as design and installation is improper, reason is not clear

7) abort situation fuzziness amb, span 0,1 two kinds

8) breakdown loss weighted value weight, span 0～1.

Distribution Network Failure is traveled through, for fault ft_i,

Case1: this location of fault type is 0, illustrates that this abort situation is precisely reliable, if its fault occurs position Put (pos_x_i,pos_y_i) ∈ g (x, y), then will record ft_iInsertion ft_list (x, y), wherein, ft_i.Amb=0, ft_i. Weight=1；

Case2: this location of fault type is 1, illustrates that this abort situation is fuzzy, comprises faulty line for all The affiliated grid of section, according to the length accounting imparting corresponding weight of each grid:

Case2.1: if two block switch ss₁、ss₂And the circuit (containing branch) between them is all located at g (x, y) Among, then will record ft_iInsertion ft_list (x, y),

Wherein, ft_i.Amb=0, ft_i.Weight=1；

Case2.2: if two block switch ss₁、ss₂And the circuit between them is (containing branch, referred to as ft_ Section) it is located at different geographic grid g (x₁,y₁)～g (x_n,y_n) in, with geographic grid g (x therein_j,y_j) as a example, will Record ft_iInsertion ft_list (x_j,y_j),

Wherein ft_i.Amb=1, ft_i.

Case3: this location of fault type is 2,

Case3.1: if pb_set and the circuit (containing branch) between them are all located among g (x, y), then will record ft_iInsertion ft_list (x, y), wherein, ft_i.Amb=0, ft_i.Weight=1；

Case3.2: if pb_set and track section (containing branch, the referred to as ft_section) position connecting these distribution transformings In different geographic grid g (x₁,y₁)～g (x_n,y_n) in.With geographic grid g (x therein_j,y_j) as a example, ft will be recorded_iInsertion ft_list(x_j,y_j),

Wherein ft_i.Amb=1, ft_i.

(3) geographic grid fault impact statistic quantification, particularly as follows:

First, select the time period [t of statistics_s,t_e] it is proposed that when a length of 1 year or more.Duration no more than 3 years because Overlong time, distribution network frame topology is it may happen that larger change

Secondly, internal fault impact in each net region is quantified.For grid g (x, y), statistical indicator is as follows:

(1b) fault frequency

$ft\_count={\σ}_{i=1}^{n^{\′}}{\mathrm{ft}}_i.weight$

Wherein, n ' is to occur Primary fault number of times in grid；

(2b) fault outage impact accumulated value

$ft\_sum=\underset{i=1}{\overset{n^{\′}}{\σ}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)$

(3b) fault outage impact line length average

$ft\_avg\_line=\frac{{\mathrm{\σ}}_{i=1}^{n^{\′}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)}{{\mathrm{\σ}}_{j=1}^{m^{\′}}(length\_grid({\mathrm{line}}_j))}$

Wherein, length_grid (line_j) it is length in grid g (x, y) for the circuit j；

Wherein, m ' is the sum of distribution transforming comprising in grid or the sum of the feeder line interlocking with grid, line_jFor in grid J-th strip feeder line；

(4b) fault outage impact 10kv capacity of distribution transform average

$ft\_avg\_tr=\frac{{\mathrm{\σ}}_{i=1}^{n^{\′}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)}{{\mathrm{\σ}}_{j=1}^{m^{\′}}\left(volumn\right({\mathrm{tr}}_j))}$

Wherein, volumn (tr_j) for distribution transformer j capacity, wherein, tr_jFor jth platform distribution transforming in grid.

It should be noted that These parameters are the overall performane regardless of fault type.Can also on the basis of failure modes, Carry out the statistics of These parameters, be beneficial to more fine accident analysis and aid decision.

(4) according to each grid fault impact statistical value generation fault impact mottle figure:

After selected statistical regions, statistical time range, fault type, statistical indicator, system generates each according to step 1～3 computing (computational methods are that point two circulations calculate each grid successively to fault statistics value two-dimensional array g_stat [x] [y] of geographic grid Numerical value, calculate g_stat [m] [n] always from g_stat [0] [0], be that here is omitted for existing method).Count To 95 probits of this array, 85 probits and 30 probits, respectively as high-risk, serious, common, more low region division (corresponding method is threshold value: statistical value is higher than 95 probits is high-risk, is serious more than 85 probits and less than 95 probits, greatly It is common in 30 probits and less than 85 probits, is relatively low less than 30 probits).Field color is defined as follows: high-risk Redness, middle danger is orange, common yellow, relatively low green.Each geographic grid is carried out according to corresponding grade Color, thus form fault impact mottle figure.When user's click geographic grid, this region is ejected with list mode and owns (or certain class) fault message, shows its details after clicking on concrete fault message.

(5) it is directed to the higher region of fault influence degree and propose corresponding aid decision suggestion

(1c) high risk zone, and it is proposed that carrying out insulating, cabling reconstructing or to carry out distribution automatic based on overhead transmission line Change upgrading；

(2c) high risk zone, and based on cable but power distribution automation low degree is it is proposed that carry out power distribution automation to circuit Upgrading；

(3c) in, high risk zone (statistical indicator is: fault outage impact line length average) is it is proposed that increase this region Line sectionalizing；

(4c) high risk zone in (fault type is: external force is destroyed) is it is proposed that strengthen the communication with unit in charge of construction, and increases this The dynamics that region is patrolled and examined；

(5c) in, high risk zone (fault type is: ageing equipment), it is proposed that strengthening this area equipment investigation, upgrades in time scarce Fall into or aging equipment；

(6c) in, high risk zone (fault type is: tree line contradiction) is it is proposed that periodically carry out this region tree trimming.

Ultimate principle and principal character and the advantages of the present invention of the present invention have been shown and described above.The technology of the industry , it should be appreciated that the present invention is not restricted to the described embodiments, the simply explanation described in above-described embodiment and description is originally for personnel The principle of invention, without departing from the spirit and scope of the present invention, the present invention also has various changes and modifications, these changes Change and improvement both falls within scope of the claimed invention.Claimed scope by appending claims and its Equivalent thereof.

Claims (7)

1. the region occurred frequently discrimination method of the Distribution Network Failure based on big data is it is characterised in that include following step:

(1) it is directed to every fault together, fuzzy diagnosis is carried out to Distribution Network Failure position based on multi-source data；Based on historical load number Carry out the assessment of Distribution Network Failure power failure load loss according to, weather data；

(2) stress and strain model is carried out to geographic view, and realize the mapping of fault and geographic grid；

(3) calculate preset time section [t_s,t_e] in each geographic grid fault impact statistical indicator, in each net region therefore Barrier impact is quantified；

(4) affect value of statistical indicant according to each grid fault geographic grid is coloured, thus generating fault impact mottle figure；

(5) region occurred frequently for fault, proposes corresponding aid decision suggestion.

2. the region occurred frequently discrimination method of the Distribution Network Failure based on big data according to claim 1 is it is characterised in that step (1), in, the method carrying out fuzzy diagnosis to described Distribution Network Failure position is as follows:

(1a) according to regular expression, the abort situation text envelope from electric network fault statistical report system, districts and cities' sole duty made a report on Breath is parsed, and identifies device numbering and device name, and searches corresponding device in production management system pms, thus really Determine abort situation (pos_x, pos_y), and labelling abort situation type is 0, if None- identified, goes to step (2a)；

(2a) the block switch remote signalling searching faulty line from distribution management system dms divides position to record, if this circuit owns Block switch is mounted with distributing automation apparatus, finds out two block switch ss of separating brake at first in line fault time range₁、 ss₂, and labelling abort situation type is 1；If circuit does not install distributing automation apparatus, go to step (3a)；

(3a) search, from power information acquisition system, the load data that fault feeder has all distribution transformers under its command, according to electric current Fall zero duration, determine the power failure duration of each distribution transforming, the distribution transforming the longest of power failure duration is added to distribution transforming set pb_set In, and labelling abort situation type is 2.

3. the region occurred frequently discrimination method of the Distribution Network Failure based on big data according to claim 2 is it is characterised in that step (2), in, the method that described geographic view is carried out with stress and strain model is as follows:

Geographic area is based on first quartile coordinate system, and the square that geographic area is len with the length of side carries out cutting it is assumed that amounting to M row n row square covers given geographic area, wherein, m > 3, n > 3, it is g (0,0) that the grid in the lower left corner is numbered, then position Numbering in the grid that xth row, y line position are put is g (x, y).

4. the region occurred frequently discrimination method of the Distribution Network Failure based on big data according to claim 3 is it is characterised in that step (2), in, described fault is as follows with the mapping method of geographic grid:

Define event of failure list ft_list (x, y) of geographic grid g (x, y), the fault occurring in this grid for storage Event, each record comprises with properties: time of failure occure_time, fault feeder feeder_info, fault Position gps coordinate x-axis pos_x, abort situation gps coordinate y-axis pos_y, fault outage loss load loss, fault type Type, abort situation fuzziness amb and breakdown loss weighted value weight；

Distribution Network Failure is traveled through, for fault ft_i,

Case1: this location of fault type is 0, if its fault occurs position (pos_x_i,pos_y_i) ∈ g (x, y), then will remember Record ft_iInsertion ft_list (x, y), wherein, ft_i.Amb=0, ft_i.Weight=1, ft_i.pos_x=pos_x_i,ft_i.pos_y =pos_y_i；

Case2: this location of fault type is 1,

Case2.1: if two block switch ss₁、ss₂And the circuit between them is all located among g (x, y), then will record ft_iInsertion ft_list (x, y), wherein, ft_i.Amb=0, ft_i.Weight=1；

Case2.2: by two block switch ss₁、ss₂And the circuit between them is referred to as ft_section, if ft_ Section is located at different geographic grid g (x₁,y₁)～g (x_n,y_n) in, fault is inserted into these ground according to different weights In the error listing of reason grid；For geographic grid g (x_j,y_j), ft will be recorded_iInsertion ft_list (x_j,y_j), wherein, ft_i. Amb=1, ft_i.

Case3: this location of fault type is 2,

Case3.1: if pb_set and the circuit between them are all located among g (x, y), then will record ft_iInsertion ft_ List (x, y), wherein, ft_i.Amb=0, ft_i.Weight=1；

Case3.2: by pb_set and the track section that connects these distribution transformings is referred to as ft_section, if ft_section position In different geographic grid g (x₁,y₁)～g (x_n,y_n) in, fault is inserted into these geographic grids according to different weights In error listing；For geographic grid g (x_j,y_j), ft will be recorded_iInsertion ft_list (x_j,y_j), wherein, ft_i.Amb=1, ft_i.

5. the region occurred frequently discrimination method of the Distribution Network Failure based on big data according to claim 4 is it is characterised in that step (3), in, described fault impact statistical indicator includes fault frequency, fault outage impact accumulated value, fault outage impact line Road length average and fault outage impact 10kv capacity of distribution transform average；

For grid g (x, y), the computational methods that described fault affects statistical indicator are as follows:

(1b) fault frequency

Wherein, n ' is the primary fault number of times occurring in grid；

(2b) fault outage impact accumulated value

$ft\_sum=\underset{i=1}{\overset{n^{\′}}{\σ}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)$

(3b) fault outage impact line length average

$ft\_avg\_line=\frac{{\mathrm{\σ}}_{i=1}^{n^{\′}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)}{{\mathrm{\σ}}_{j=1}^{m^{\′}}(length\_grid({\mathrm{line}}_j))}$

Wherein, length_grid (line_j) it is length in grid g (x, y) for the circuit j；

M ' is the sum of distribution transforming comprising in grid or the sum of the feeder line interlocking with grid, line_jFor j-th strip feeder line in grid；

(4b) fault outage impact 10kv capacity of distribution transform average

$ft\_avg\_tr=\frac{{\mathrm{\σ}}_{i=1}^{n^{\′}}({\mathrm{ft}}_i.loss*{\mathrm{ft}}_i.weight)}{{\mathrm{\σ}}_{j=1}^{m^{\′}}\left(volumn\right({\mathrm{tr}}_j))}$

Wherein, volumn (tr_j) for distribution transformer j capacity, wherein, tr_jFor jth platform distribution transforming in grid.

6. the region occurred frequently discrimination method of the Distribution Network Failure based on big data according to claim 1 is it is characterised in that step (4), in, described fault impact mottle map generalization method is as follows:

After selected statistical regions, statistical time range, fault type, statistical indicator, calculate the fault statistics value two dimension of each geographic grid Array g_stat [x] [y]；

Statistics obtains 95 probits of this array, 85 probits and 30 probits, respectively as high-risk, serious, common and relatively low The threshold value of region division；

Each geographic grid is coloured according to corresponding grade, thus forming fault impact mottle figure.

7. the region occurred frequently discrimination method of the Distribution Network Failure based on big data according to claim 1 is it is characterised in that step (5), in, described aid decision suggestion is as follows:

(1c) high risk zone, and based on overhead transmission line, then suggestion carries out insulating, cabling reconstructing or carries out power distribution automation Upgrading；

(2c) high risk zone, and based on cable but power distribution automation low degree, then advise carrying out power distribution automation liter to circuit Level transformation；

(3c) high risk zone in, and statistical indicator affects line length average for fault outage, then suggestion increases the line in this region Road segmentation；

(4c) high risk zone in, and fault type is external force destruction, then the communication with unit in charge of construction is strengthened in suggestion, and increases this area The dynamics that domain is patrolled and examined；

(5c) high risk zone in, and fault type is ageing equipment, then this area equipment investigation is strengthened in suggestion, and upgrade in time defect Or aging equipment；

(6c) high risk zone in, and fault type is tree line contradiction, then advise periodically carrying out this region tree trimming.