CN106019047A - Comprehensive line selection fault measurement calculation method for small-current grounding system - Google Patents

Abstract

The invention discloses a comprehensive line selection fault measurement calculation method for a small-current grounding system, and the method comprises the following steps: starting a fault line selection logic; collecting fault data; selecting M line selection criteria and N fault measurement calculation methods; calculating the fault measurement value of each line (including the bus) for the selected line selection criterion through employing the selected fault measurement calculation method, wherein each line has N fault measurement values; and integrating the N fault measurement values of each line as the fault measurement value of the line for the selected line selection criterion through employing the DS evidence theory. According to the invention, the method employs the DS evidence for integrating conventional different fault measurement calculation methods, overcomes the one-sidedness of the conventional different fault measurement calculation methods, further improves the fault measurement calculation accuracy after integration, and improves the accuracy of fault line selection.

Description

Small current neutral grounding system faulty line selection fault measurement computational methods

Technical field

The present invention relates to single-phase ground fault line selecting method of small-electric current grounding system, especially relate to a kind of small current neutral grounding The computational methods of fault measurement in system single-phase earth fault faulty line selection.

Background technology

In China, most 3～66kV medium voltage distribution networks all use isolated neutral or through grounding through arc The method of operation, referred to as system with non effectively earth ed neutral.When after system with non effectively earth ed neutral generation singlephase earth fault, by In neutral non-effective grounding, circuit does not constitute loop, therefore circuit does not has short circuit current to flow through, simply via line pair Ground electric capacity forms less current path, therefore, is small current neutral grounding system also known as system with non effectively earth ed neutral.Due to this Time three-phase line voltage remain in that symmetrical relations, do not affect power supply continuous to load, therefore need not trip immediately, code regulation permissible Continue to run with 1～2 hour.Although the singlephase earth fault that small current neutral grounding system occurs does not interferes with the properly functioning of electrical network, But owing to now healthy phases voltage-to-ground raises, it is susceptible to healthy phases flashover over the ground, thus causes two-phase grounding fault, It is also possible to cause the accident such as voltage transformer and bus burns, electric cable explosion, Power Plant stoppage in transit, it is therefore desirable at short notice Correctly select faulty line, in order to operations staff adopts an effective measure in time and deals carefully with.

Due to faint by the complexity of electrical network, fault-signal and there is the factors such as various interference and affected, small area analysis The integrity problem of single-phase grounded malfunction in grounded system route selection becomes the difficult problem perplexing the work of domestic and international relay protection for a long time, Up to now, Chinese scholars proposes substantial amounts of selection method, wherein uses the method for information fusion to carry out multiple criterions The faulty line selection merged is the means of a kind of reliability being effectively improved failure line selection.Commonly used fault in faulty line selection method Estimating this concept, fault measurement is defined in the real variable in a certain interval, is used for describing a circuit under certain criterion The sizes values of the doubtful earth fault line information characteristics degree comprised.When the fault measurement of a certain bar circuit is the closer to being determined The real variable interval maximum of justice, shows that this circuit is that faulty line probability is the highest.Whereas if the fault of certain circuit is surveyed Degree, closer to defined real variable interval minima, shows that this circuit is that non-fault line probability is the highest.Presently, there are many Plant the computational methods of fault measurement, Jia Qingquan；Yang Qixun etc. are at the periodical " Proceedings of the CSEE " that number of the edition is 0258-8013 Page 6 to page 11 " power distribution network single-phase earthing event based on fault measurement concept with evidence theory of the 12nd phase of volume 23 in 2003 Hinder many criterions to merge " disclose a kind of fault measurement computational methods, write at Pang Qingle, Electronic Industry Press in August, 2010 The monograph " low-current ground fault line selection and location technology " page 85 to page 93 published discloses a kind of fault measurement meter Calculation method, Kang Yi, Liu Peiyue etc. are in periodical " North China Electric Power University's journal " the 3rd phase of volume 34 in 2007 of number of the edition 1007-2691 Page 6 to page 11 " application fuzzy theory realizes one-phase earthing failure in electric distribution network and differentiates " discloses a kind of fault measurement calculating side Method, owing to the computational methods of the above fault measurement are all that each scholar is defined according to the understanding of self and designs, There is certain one-sidedness, and final faulty line selection result is had a great impact by the accuracy of fault measurement, therefore, on State these fault measurement computational methods and affect the raising of faulty line selection accuracy owing to there is one-sidedness.

Summary of the invention

The technical problem to be solved in the present invention is to provide the calculating of a kind of small current neutral grounding system faulty line selection fault measurement Method, this approach reduces the one-sidedness of existing fault measurement computational methods, improves the accuracy of fault measurement, thus Improve the accuracy of faulty line selection.

For solving above-mentioned technical problem, the present invention uses the method for DS evidence theory by current existing several major failures Likelihood Computation result carries out merging the fault measurement result as last, thus reduces the one-sidedness of fault measurement computational methods, Improve the accuracy of fault measurement, contribute to improving faulty line selection accuracy further.

The technical scheme that this method is taked comprises the following steps:

Step 1, failure line selection logic start；

Step 2, collection fault data；

Step 3, M route selection criterion of selection and N kind fault measurement computational methods；

Step 4, choose a route selection criterion；

The N kind fault measurement computational methods that step 5, utilization are chosen calculate every line fault measure value, then each line Rood is to N number of fault measurement；

Step 6, for each circuit, the N number of fault measurement utilizing DS evidence theory fusion step 5 to obtain is sentenced as this According to the fault measurement to every circuit；

Step 7, choose next route selection criterion, repeat the 5th step and the 6th step, until m-th criterion calculates complete.

Information fusion is exactly under certain criterion, uses corresponding theoretical and algorithm, utilizes computer to enter multi-source information Row processes, it is achieved by acquisition information carry out comprehensively utilizing thus reach more can accurately reflect the purpose of practical situation.At source number Inevitably there is various noise according to, use different information to describe the things of same feature, it is possible to reduce this By the uncertainty caused by source data inaccuracy, the precision of system can be greatly improved.Information fusion technology makes purpose Justice shows themselves in that higher information utilization, higher fault-tolerant ability, higher precision, higher detection and identification ability.DS demonstrate,proves Being information fusion algorithm based on statistics according to theory, the feature of its maximum is that the description to uncertain information have employed that " interval is estimated Meter " rather than the method for " point estimation ", do not know to show in terms of uncertain aspect and accurate reflection evidence-gathering in differentiation Go out the biggest motility.

The technology of the present invention effect is:

The computational methods of current fault measurement can regard the measurement to real fault measurement as, owing to they all have Having certain one-sidedness, therefore they are the measured values that there is noise, use DS evidence theory various faults to be estimated and melt Close, reach the purpose using different information to describe fault measurement, it is possible to reduce this by caused by source data inaccuracy Uncertainty, improves the accuracy measured.Therefore the fault measurement accuracy after using DS evidence theory fusion improves further, Thus improve the accuracy of failure line selection.

Accompanying drawing explanation

Fig. 1 is the flow chart that the present invention realizes；

Fig. 2 is that earth fault of the present invention tests schematic diagram.

Detailed description of the invention

Below in conjunction with the accompanying drawings 1 and example invention is described further, the present invention specifically comprises the steps of

Step 1, failure line selection logic start: when systems are functioning properly, small current neutral grounding system residual voltage is zero, three Phase phase voltage balances, and when there is singlephase earth fault, its residual voltage is more than zero, and faulted phase voltage reduces, healthy phases electricity Pressure raises, and therefore, it can increase to over setting value with bus residual voltage instantaneous value more than setting value and healthy phases voltage, Judging, when condition is satisfied by, fault line selection device starts；Wherein residual voltage setting value is the phase voltage of 0.15 times, Healthy phases voltage setting value is 1.1 times of phase voltages；

After step 2, line selection apparatus start, gather 15 cycle data after front 2 cycles of fault and fault, before extracting fault Half cycle and fault later half cycles data totally two cycle data, as transient data, extract after fault 5～15 Cycle data are as steady-state quantity data；

Step 3, M route selection criterion of selection and N kind fault measurement computational methods: selection 1～2 kind of steady-state criterion and 2～3 kind Transient state criterion is as the criterion of faulty line selection, selection 3～4 kind of existing conventional fault measurement computational methods；

Step 4, choose a route selection criterion；

Step 5, for a selected route selection criterion, calculate the fault measurement under every kind of fault measurement computational methods, then For this route selection criterion, every circuit has N kind fault measurement；

Step 6, the N kind fault measurement utilizing DS Method of Evidence Theory to obtain step 5 merge, and obtain after fusion Fault measurement is then every line fault measure value under this criterion.

Specifically comprising the following steps that of DS evidence theory fusion

6A, the determination of identification framework

The determination of fault measurement is primarily directed to the fault measurement of all circuits of system, and therefore, all circuits of system are with regard to structure Become the identification framework of fault measurement.

If system has L bar circuit, compile a numbering to every circuit, represent by array S (), then identification framework is Θ =S (k) | k=0,1 ..., L}

6B, construct basic brief inference function

Basic brief inference function is defined as identification framework Θ power set 2^ΘOn function, if set function m:2^Θ→ [0,1] Meet m (φ)=0；ForThen m (A) is referred to as the basic brief inference function of A.M (A) reflects Size to the reliability of A itself.

In the case of the failure criterion determined, the computational methods of each fault measurement just constitute an evidence body, For every kind of criterion, select M kind fault measurement computational methods, then can get M kind fault measurement.

The basic brief inference function that the present invention builds is:

M=m_r·m_a (1)

M in formula_rFor relative brief inference function, m_aFor can determine that reliability coefficient.

For kth bar circuit, in the case of the failure criterion determined, for a kind of fault measurement calculating side determined Method.Its relative brief inference function is calculated as follows:

First fault measurement and all line fault of system calculating kth bar circuit estimates the ratio of sum:

$x\left(k\right)=F\left(k\right)/\underset{s=1}{\overset{L}{\mathrm{\Σ}}}F\left(k\right)---\left(2\right)$

Wherein F (k) represents the fault measurement of kth bar circuit.

Then kth bar circuit calculates according to following piecewise function relative to brief inference function:

$m_r=\left\{\begin{array}{cc}0,& x\&le;0.15\\ 2(x-0.15),& 0.15<x<0.65\\ 1,& x\&GreaterEqual;0.65\end{array}\right.---\left(3\right)$

Assume for M kind fault measurement computational methods, z_i, i=1 ..., M is that i-th kind of fault measurement computational methods obtains Including the maximum fault measurement value of all circuits of bus, z_mIt is z_iMiddle maximum, then can determine that reliability coefficient is

m_a=z_i/z_m (4)

Then basic brief inference function expression is:

M (A)=m_r(A)·m_aA={S (k) } (5)

$m\left(\mathrm{\&Theta;}\right)=1-\underset{A\&Subset;\mathrm{\&Theta;}}{\mathrm{\&Sigma;}}m\left(A\right)---\left(6\right)$

Wherein uncertain trusting degree is distributed in m (Θ) expression.

6C, Evidence Combination Methods

For selected criterion, it is assumed that have employed N kind fault measurement computational methods, h_j(j=0 ..., L) represent that circuit j is Faulty line, available N group basic brief inference value:

m₁(h₀),m₁(h₁),m₁(h₂),…,m₁(h_L),m₁(Θ)；

m₂(h₀),m₂(h₁),m₂(h₂),…,m₂(h_L),m₂(Θ)；

…………

m_N(h₀),m_N(h₁),m_N(h₂),…,m_N(h_L),m_N(Θ)； (7)

DS evidence theory rule of combination is:

1) composition rule of two brief inference numbers

If M₁And M₂It is 2^ΘUpper two brief inference functions, then its orthogonal and M=M₁+M₂It is defined as:

M (Φ)=0, A=Φ

$M\left(A\right)=c^{-1}\underset{A_1\&cap;A_2=A}{\mathrm{\&Sigma;}}{M}_1\left(A_1\right)M_2\left(A_2\right),A\&NotEqual;\mathrm{\&Phi;}---\left(8\right)$

Wherein

$c=1-\underset{A_1\&cap;A_2=\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}{M}_1\left(A_1\right)M_2\left(A_2\right)=\underset{A_1\&cap;A_2\&NotEqual;\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}{M}_1\left(A_1\right)M_2\left(A_2\right)---\left(9\right)$

In formula, c is to comprise conflict completely to assume A₁And A₂All brief inference function product sums, A is assumed in so-called conflict₁ And A₂Refer to that target pattern can not exist in identification framework Θ simultaneously, refer to circuit h in the present invention_j(j=0 ..., L) Can not break down, therefore, composition rule algorithm concrete in the present invention is as follows simultaneously:

$\begin{array}{c}c=1-\underset{h_i\&cap;h_j=\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}{m}_1\left(h_i\right)m_2\left(h_j\right)=1-\underset{i=0}{\overset{L}{\mathrm{\&Sigma;}}}\&lsqb;m_1\left(h_i\right)\&CenterDot;\underset{j\&NotEqual;i}{\overset{L}{\underset{j=0}{\mathrm{\&Sigma;}}}}{m}_2\left(h_j\right)\&rsqb;\\ =1-\{m_1\left(h_0\right)\&lsqb;m_2\left(h_1\right)+m_2\left(h_2\right)+\mathrm{...}+m_2\left(h_L\right)\&rsqb;+\mathrm{...}+m_1\left(h_L\right)\&lsqb;m_2\left(h_0\right)+m_2\left(h_1\right)+\mathrm{...}+m_2\left(h_{L-1}\right)\&rsqb;\}\end{array}---\left(10\right)$

$\underset{h_i\&cap;h_j=h_0}{\mathrm{\&Sigma;}}{m}_1\left(h_i\right)m_2\left(h_j\right)=m_1\left(h_0\right)m_2\left(h_0\right)+m_1\left(h_0\right)m_2\left(h_{\mathrm{\&Theta;}}\right)+m_1\left(h_{\mathrm{\&Theta;}}\right)m_2\left(h_0\right)---\left(11\right)$

Then can be assigned to circuit h after merging₀Brief inference functional value

$m\left(h_0\right)=c^{-1}\underset{h_i\&cap;h_j=h_0}{\mathrm{\&Sigma;}}{m}_1\left(h_i\right)m_2\left(h_j\right)---\left(12\right)$

The brief inference functional value m (h of All other routes it is assigned to after in like manner can obtaining fusion_i), i=1 ..., L

Uncertainty after fusion

M (Θ)=m₁(Θ)m₂(Θ)/c (13)

2) composition rule of multiple brief inference functions

Orthogonal and the M=M of multiple brief inference functions₁+M₂+…+M_nIt is defined as:

M (Φ)=0, A=Φ

$M\left(A\right)=c^{-1}\underset{\&cap;A_i=A}{\mathrm{\&Sigma;}}\underset{1\&le;i\&le;n}{\mathrm{\&Pi;}}{M}_i\left(A_i\right),A\&NotEqual;\mathrm{\&Phi;}---\left(14\right)$

Wherein

$c=1-\underset{\&cap;A_i=\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}\underset{1\&le;i\&le;n}{\mathrm{\&Pi;}}{M}_i\left(A_i\right)=\underset{\&cap;A_i\&NotEqual;\mathrm{\&Phi;}}{\mathrm{\&Sigma;}}\underset{1\&le;i\&le;n}{\mathrm{\&Pi;}}{M}_i\left(A_i\right)---\left(15\right)$

Multiple brief inference functions composition rule in the present invention is specifically calculated, its method and two brief inference The composition rule specific algorithm of function is similar to.

6D, the determination of fault measurement

For the criterion determined, using the result of DS Evidence Combination Methods as the fault measurement value of every circuit；

Step 7, choose next route selection criterion, repeated execution of steps 5 and step 6, until m-th criterion.

Specific embodiment:

Fig. 2 is single-phase grounded malfunction in grounded system of low current schematic diagram, and in figure, G is power supply, and T is transformator, and no-load voltage ratio is 110kV/11kV, connecting group grade is YN/D11, has 5 outlets, is isolated neutral system when switching K and disconnecting, as switch K During Guan Bi be neutral point through arc suppression coil series resistance ground connection, overcompensation degree is 10%, and the positive order parameter of circuit is: R₁=0.17 Ω/km, L₁=7.6mH/km, C₁=0.061uF/km, Zero sequence parameter is: R₀=0.23 Ω/km, L₀=34.4mH/km, C₀= 0.038uF/km.The length of circuit is respectively as follows: L₁=12km, L₂=20km, L₃=25km, L₄=6km, L₅=22km.

Respectively at the diverse location of every circuit (comprising bus), with different fault close angles, different voltage-phases And under the conditions of different transition resistance etc., be provided with substantial amounts of trouble point and carry out simulating, verifying.For isolated neutral system, Criterion chooses zero-sequence current fundametal compoment method, zero-sequence current quintuple harmonics component method and wavelet packet method these three criterion, for Neutral by arc extinction coil grounding system, criterion is chosen zero-sequence current real component method, wavelet packet method and suddenlys change mensuration.Fault Likelihood Computation method is respectively adopted document 1: " Proceedings of the CSEE " the 12nd phase in 2003 is " based on fault measurement concept and card Merge according to the theoretical many criterions of one-phase earthing failure in electric distribution network " disclosed in fault measurement computational methods, document 2: Pang Qingle writes, Fault measurement disclosed in the monograph " low-current ground fault line selection and location technology " that Electronic Industry Press publishes in August, 2010 Computational methods, document 3: periodical " North China Electric Power University's journal " the 3rd phase in 2007, " power distribution network is single-phase to be connect to apply fuzzy theory to realize Earth fault differentiate " disclosed in fault measurement computational methods, with the present invention utilize DS evidence theory by three of the above document calculate The fault measurement that the fault measurement gone out obtains after merging contrasts.Table 1 is several after system every circuit arranges various faults Method fault measurement value of calculation be averaged after results contrast.Vector, (x in table₁,x₂,x₃,x₄,x₅,x_s) represent circuit respectively L₁、L₂、L₃、L₄、L₅, the fault measurement value of bus, it is interval that fault measurement value all normalizes to [0,1].

The mean failure rate Likelihood Computation results contrast of the various method of table 1

As it can be seen from table 1 use the fault measurement of the present invention, it is possible to faulty line is highlighted with becoming apparent from, It is thus possible to improve the accuracy of faulty line selection further.

Table 2 is to utilize above-mentioned fault measurement to use neutral net faulty line selection method to carry out the accuracy of route selection.

Table 2 is estimated based on different faults and is carried out the accuracy of neutral net route selection and compare

By table 2, the method utilizing several fault measurement to estimate all can effectively select faulty line, but utilizes this Invention fault measurement can improve route selection accuracy higher.

Claims (2)

1. small current neutral grounding system faulty line selection fault measurement computational methods, it is achieved process be:

(1) failure line selection logic starts；

(2) fault data is gathered；

(3) route selection criterion and fault measurement computational methods are selected；

(4) every including the bus line fault measure value that the route selection criterion chosen gives is calculated；

It is characterized in that:

(5), for the route selection criterion chosen, when calculating every line fault measure value, various faults hygrometer is calculated Value merge after result as the fault measurement value of every circuit.

2., according to the small current neutral grounding system faulty line selection fault measurement computational methods described in right 1, it is characterized in that: selected Fusion method is DS evidence theory.