CN105866633A - Power transmission line fault current traveling-wave waveform recurrence method based on wave weights - Google Patents

Abstract

The invention discloses a power transmission line fault current traveling-wave waveform recurrence method based on wave weights. A fault traveling-wave detecting device is installed on a power transmission line, a frequency-dependent model based objective function is firstly set by adopting a least square method according to fault point current traveling waves detected by two of points, and distribution parameters of the power transmission line are inverted out by adopting an intelligent search algorithm simulated annealing algorithm. The correctness of the distribution parameters is verified by inverting waveforms of detection points of another section of the same line and comparing the waveforms with really detected waveforms, and finally fault current traveling-wave waveforms of any known points on a faulted line can be accurately inverted by adopting the inversion method. An absolute error of the method does not exceed 3 microseconds, and a relative error does not exceed 6%.

Description

The replay method of transmission line malfunction current traveling wave waveform based on ripple weight

Technical field

The present invention relates to transmission line malfunction testing techniques of equipment field, specifically a kind of based on ripple The replay method of the transmission line malfunction current traveling wave waveform of weight.

Background technology

Transmission line of electricity is the equipment the most easily broken down in power system, for the event of transmission line of electricity Barrier range finding and fault location, scholar has carried out many research.For the fault of transmission line of electricity, The most passively increase its defence capability to reduce the probability that transmission line of electricity breaks down.And in reality During circuit local flow improvement engineer applied, owing to lacking fault data as support so that transformation work Work cannot be accomplished with a definite target in view, thus wastes the investment of substantial amounts of manpower and materials.If fault can be recurred Evolution, it is achieved transmission line malfunction whole status of processes sight is visual, to power transmission line The side such as road accident analysis, diagnosis, the formulation of the circuit defense schemes in later stage and failure accident manoeuvre There is major and immediate significance in face.

For uniline, if it is uniform, unified to set this section of transmission line of electricity, the electricity of its unit length Resistance, inductance, electric capacity and conductance are respectively R, L, G, C, and transmission line of electricity takes a segment length For the circuit of dx, this circuit propagation equation in frequency domain is:

$\left\{\begin{array}{c}-\frac{dU}{dx}=(R+j\mathrm{\ω}L)I\\ -\frac{dI}{dx}=(G+j\mathrm{\ω}C)U\end{array}\right.---\left(1\right)$

The solution of formula (1) may finally be written as form:

$\{\begin{array}{c}{U}_x=A_1{e}^{-\mathrm{\λ}x}+A_2{e}^{\mathrm{\λ}x}\\ I_x=A_1{e}^{-\mathrm{\λ}x}/Z_c-A_2{e}^{\mathrm{\λ}x}/Z_c\end{array}---\left(2\right)$

Wherein,For line propagation coefficient, x is propagation distance, Z_cFor natural impedance.A₁、A₂For the integral constant determined by boundary condition.

For single fault traveling wave, if being not required to the anti-row ripple considering to propagate along x opposite direction, formula (2) Can be written as:

$\left\{\begin{array}{c}{U}_x^{*}=A_1^{*}{e}^{-{\mathrm{\λ}}^{*}x}\\ I_x^{*}=A_1^{*}{e}^{-{\mathrm{\λ}}^{*}x}/Z_c^{*}\end{array}\right.---\left(3\right)$

Can be obtained by formula (3), on circuit at a distance of for the 2: 1 and 2 of x, between them Current wave and voltage wave have following relation

$\left\{\begin{array}{c}{U}_{x_2}=U_{x_1}{e}^{-\mathrm{\λ}(x_2-x_1)}=U_{x_1}{e}^{-\mathrm{\λ}x}\\ I_{x_2}=I_{x_1}{e}^{-\mathrm{\λ}(x_2-x_1)}=I_{x_1}{e}^{-\mathrm{\λ}x}\end{array}\right.---\left(4\right)$

By formula (4) it can be seen that ripple is close with its frequency in the propagation of single phase homogeneous transmission line of electricity Cut is closed, and along with the increase of propagation distance x, voltage, electric current are gradually decayed.Note H=e^-λx The transmission function propagated along the line for transmission line of electricity frequency domain, then this propagation model is referred to as transmission line of electricity Frequency is according to function model.

And for three phase line, the three alternate coupled relations that exist, need to be by phase moding Transformation becomes three independent components, in order to analyze.

For on circuit at a distance of for the 2: 1 and 2 of d, by (4) formula after phase-model transformation Finally give following current traveling wave waveform relationship

$\left\{\begin{array}{c}{I}_2^{\left(i\right)}=I_1^{\left(i\right)}{H}^{\left(i\right)}=I_1^{\left(i\right)}{e}^{-{\mathrm{\λ}}^{\left(i\right)}d}\\ {\mathrm{\λ}}^{\left(i\right)}=\sqrt{(R^{\left(i\right)}+{\mathrm{j\ωL}}^{\left(i\right)})(G^{\left(i\right)}+{\mathrm{j\ωC}}^{\left(i\right)})}\end{array}\right.---\left(5\right)$

Wherein, subscript i (i=0,1,2) represents i mold component.

If wondering the transportation law of transmission line of electricity, four distributed constants in λ must it is known that And from formula (5) it will be seen that transmission function λ have four variablees and the impact that intercouples, Traditional mathematical method solves for this class equation and has become unable to do what one wishes, and hardly results in Globally optimal solution.Therefore need to use global optimization intelligent algorithm simulated annealing.

Simulated annealing (Simulated Annealing, SA) thought the earliest is by N. Metropolis et al. proposed in nineteen fifty-three.Nineteen eighty-three, S.Kirkpatrick etc. successfully will move back Fire thought is incorporated into Combinatorial Optimization field.It is based on Monte-Carlo iterative strategy A kind of random optimizing algorithm, its starting point is based on the annealing process of solid matter and in physics As similarity between combinatorial optimization problem.Simulated annealing from a certain higher initial temperature, With the continuous decline of temperature parameter, join probability kick characteristic finds mesh in solution space at random The globally optimal solution of scalar functions, i.e. probability can jump out and finally tend to complete at locally optimal solution Office is optimum.

In terms of fault recurrence, the fault recurrence of Chinese scholars more attention location system aspect, As North China Electric Power University Zhang Dongying relies on model of electrical network, electrical network real time execution, failure wave-recording It is true that the information such as information, protection act information and failure process breaker in middle action first build fault zone Determine suspect device, then actively collect the information relevant to suspect device according to suspect device, combine Conjunction utilizes failure wave-recording result as middle junction opinion, finally utilizes evidence theory, positive and negative mixing to push away The methods such as reason finally determine faulty equipment and tentatively judge failure process.But for defeated More utilizations of line fault traveling-wave waveform information and data mining do not have fully excavation.

Summary of the invention

The disappearance of this block technology is recurred in order to solve trouble point information, it is desirable to provide a kind of The replay method of transmission line malfunction current traveling wave waveform based on ripple weight, examines along the line with fault The row wave datum that measuring point detects is foundation, is finally inversed by defeated by intelligent algorithm simulated annealing Distributed constant in electric line, goes out circuit trouble point in conjunction with transmission line of electricity frequency according to model inversion Traveling-wave waveform.

The technical solution of the present invention is as follows:

A kind of replay method of transmission line malfunction current traveling wave waveform based on ripple weight, the party Method comprises the following steps:

Step S1: set multiple fault detecting point on transmission line of electricity successively, for fault electricity The collection of popular waveform；

Step S2: distance fault on transmission line of electricity is actually occurred place's homonymy except a nearest inspection The waveform of the another two fault detecting point of measuring point carries out card human relations boolean and converts three-phase decoupling；

Step S3: carry out after each modulus of row ripple after three-phase decouples is carried out wavelet package transforms Fast Fourier transform, point frequency band substitutes into circuit frequency according to function model, concrete steps:

Step S3.1: as a example by N shell wavelet package transforms, for inciting somebody to action after N shell wavelet package transforms Two test points be respectively divided into 2^NEach frequency range waveform in individual frequency range, is carried out in quick Fu Leaf transformation generates respective 2^NSection frequency domain data.

Step S3.2: will be close to the frequency domain data of each modulus of test point at that of trouble point and be multiplied by Circuit according to frequency function H, and the data of the corresponding frequency band of modulus corresponding with another test point do poor Δ d (i), i represents i-th frequency range；

Wherein H=e^-λx, x is the distance between two test points, For each frequency range, f is the mid frequency of this frequency range, and R is resistance, and L is inductance, and G is electricity Leading, C is electric capacity.

Step S4: each frequency band is passed through the wave type energy of the detection waveform near trouble point Accounting is weighted, as coefficient, the target optimizing function that combination producing is final；

Step S4.1: as a example by N shell wavelet package transforms, to the test point pointed out near fault Decoupling waveform carry out card human relations boolean and convert decoupling.

Step S4.2: each modulus waveform after decoupling is carried out N shell wavelet package transforms and is divided into 2^N Individual frequency range.

Step S4.3: the wave type energy that the component of each frequency range is had accounts for the hundred of total wave type energy Proportion by subtraction is exactly weight a (i) of weighted array, and i represents i-th frequency range；

Step S4.4: the object function after final weighting is as follows:* Represent conjugate transpose.

Step S5: by simulated annealing, target optimizing function is carried out the overall situation and solve, Distributed constant to this section of transmission line of electricity；

Step S6: the distributed constant that obtains is substituted into circuit frequency according to function model, according to known Point waveform combined circuit frequency is finally inversed by unknown point waveform according to parameter model, concrete steps:

S6.1: after obtaining circuit distributed constant in S5, determines that circuit frequency is according to function model H, wherein H=e^-λx, x is the distance of traveling wave, is entered by the known point waveform being used for inverting Row card human relations boolean converts decoupling becomes each modulus waveform；

S6.2: the traveling wave line mould waveform after decoupling three-phase carries out fast Fourier transform, according to The position relationship of unknown point and known point and circuit frequency do computing according to function model:

If unknown point is in the downstream of known point, then known point each modulus waveform is multiplied by circuit frequency according to letter Digital-to-analogue type H；

If unknown point is in the middle of known point and line fault point, then known point each modulus waveform divided by Circuit frequency is according to function model H；

S6.3: according to the line mould waveform of unknown point after computing, in conjunction with fault type boundary condition, Push away to obtain zero mould waveform of unknown point；

S6.4: carry out card human relations boolean's inverse transformation according to each modulus waveform of unknown point and obtain unknown point Three-phase fault current traveling wave waveform.

Distributed constant inversion principle be based on the row ripple making row wave loops model theory push away and The residual error of the actual waveform that observation measurement obtains minimizes and carries out object function optimizing, this It is least square object function used by bright:

Q=Δ d Δ d^* (6)

Δ d=d in formula_cal-d_obs, d_calBe current TRANSFER MODEL corresponding just drill data, d_obsIt is actual Measuring the data arrived, * represents conjugate transpose.

And the current traveling wave waveform of analysis of the present invention is a high frequency transient signal, for its details During extraction, we have used wavelet package transforms signal have carried out frequency-division section process, thus finally Object function is to be formed by combined by the component of each frequency band, and form is as follows:

$Q=\underset{i=1}{\overset{n}{\Σ}}a\left(i\right)\·\mathrm{\Δ}d\·{\mathrm{\Δd}}^{*}---\left(7\right)$

The weight coefficient of i-th section after a (i) represents (n-1)th layer of wavelet package transforms in formula, by ripple The energy of each frequency range of shape accounts for the percentage ratio of gross energy and determines；D is through card human relations boolean's three-phase solution Wave data after coupling fast Fourier transform.

Seek by the object function intelligent algorithm simulated annealing in (7) formula is carried out the overall situation Excellent, obtain on circuit each modulus distributed constant of circuit between two test points.

Obtain on circuit after the distributed constant of each modulus, transfer function H just it has been determined that from And frequency can be combined according to the waveform of known test point and calculate other points on circuit according to model H Waveform and compare with the true detection waveform of this point, verify inverted parameters correctness and The feasibility of waveform inversion method.

The present invention utilizes row ripple frequency in transmission line of electricity to build fault current ripple according to propagation characteristic The inverse model of shape, by asking for circuit distributed constant and fault current traveling-wave waveform anti- Drilling, technique effect is as follows:

1) by building the inverse model of fault current traveling-wave waveform, in conjunction with wavelet package transforms with fast Speed Fourier transformation sets up least square object function, and by intelligent search algorithm simulated annealing Algorithm global optimizing, can efficiently and accurately obtain on transmission line of electricity according to two test point waveforms Distributed constant.

2) circuit distributed constant is first sought, can be according to the waveform of known point in conjunction with row wave loops model Exact inversion goes out the waveform of unknown point.

3) the row ripple of each point on circuit can be accurately obtained, utilize for the later stage traveling wave fault to believe comprehensively Breath, carries out accident analysis and does technical support.

Accompanying drawing explanation

Fig. 1 is transmission line of electricity pscad model schematic

Fig. 2 is test point 1,2,3 current traveling wave waveform

Fig. 3 is each modulus oscillogram for parametric inversion

Fig. 4 is the current traveling wave comparison of wave shape figure of test point 1

Detailed description of the invention

Below in conjunction with the accompanying drawings, provide presently preferred embodiments of the present invention, and be described in detail.

PSCAD sets up such as the model of power transmission system of Fig. 1.Circuit uses frequency according to model, F For lightning failure point, in trouble point, 20km, 40km and 70km along the line set up three detections successively Point, for the collection of fault current traveling-wave waveform.The shaft tower of transmission line of electricity is actual shaft tower ZB1 Model.

Circuit was struck by lightning at a F when 0.2 second, with the arteries and veins of a high frequency in pscad Rush signal to be simulated, produce the transient current travelling waves propagated along the line, at detected downstream point 1,2,3 The current traveling wave waveform after decay and distortion detected respectively as in figure 2 it is shown, the adopting of traveling-wave waveform Sample frequency takes 1MHz.

As seen from Figure 2, after F point breaks down, test point 1,2, first electricity of 3 Popular waveform wave head portion big with wave rear is the most relative intact, but owing to circuit row ripple exists catadioptric Penetrate, therefore the single complete waveform obtaining first ripple cannot be obtained.So that inverting is more accurate, Intend clipping first wave head and second wave head lap herein, and it is carried out card human relations boolean change Changing three-phase decoupling, each modulus waveform after decoupling is as shown in Figure 3

The each modulus of row ripple after three-phase decouples is carried out 5 layers of wavelet package transforms and is divided into 32 frequency ranges, Row of going forward side by side carries out fast Fourier transform, and point frequency band substitutes into circuit frequency according to function model H.

To each frequency band by the detection waveform (being test point 2) near trouble point herein The target optimizing function that wave type energy accounting is weighted combination producing final as coefficient is as follows:

$Q=\underset{i=1}{\overset{32}{\Σ}}a\left(i\right)\·\mathrm{\Δ}d\left(i\right)\·\mathrm{\Δ}d{\left(i\right)}^{*}$

By simulated annealing, target optimizing function is carried out the overall situation to solve, obtain this section of transmission of electricity Each modulus distributed constant of circuit is as shown in table 1:

Distributed constant between table 1 test point 2,3

For the effect of waveform inversion, the present invention is evaluated with following 5 aspects.

1) wave head initial time t_s。

2) the wave head rise timeI.e. in the amplitude of ripple from the 0.1 of maximum amplitude Rise to 0.9 times of time used again.

3) fault traveling wave peak I_m。

4) the position t of peak value_m。

5) half-wave length t_hm=t_h-t_s.Wherein t_hIncrease to for amplitude and the highest drop to half again Time moment.

By the comparative evaluation of above 5 aspects, each side characteristic of the omnibearing reflected waveform of energy.

Owing to the inverting of zero _exit is restricted by each side, can not accurately inverting, therefore this Invention abandons the inverting to zero mould, then by zero mould during different faults type and the limit of line mould Boundary's condition, the inversion result of joint line mould obtains zero _exit.This example fault type is that A connects Ground, can be in the hope of circuit inverting zero _exit by boundary condition and phase mould relation.Again to each mould Amount waveform carries out card human relations boolean's inverse transformation and obtains three-phase current comparison of wave shape figure as shown in Figure 4.

Data in Fig. 4 first wave head, 5 appraisement systems of the present invention are evaluated as follows:

1) wave head initial time t_s.Three-phase inverting waveform and detection waveform absolute error are respectively A 0.5 μ s mutually, B phase 1 μ s, C phase 2 μ s.

2) the wave head rise timeThe relative error of A, B, C three-phase is respectively 1.698%, 3.425%, 4.875%.

3) fault traveling wave peak I_m.The relative error of A, B, C three-phase is respectively 1.774%, 1.852%, 4.691%.

4) the position t of peak value_m.The absolute error of A, B, C three-phase 2 μ s, 1 μ s, 1 μ s respectively.

5) half-wave length t_hm=t_h-t_s.The relative error of A, B, C three-phase is respectively 0.723%, 2.820%, 4.669%.

Through a large amount of simulation results shows, five are commented by this fault current traveling-wave waveform inversion method Valency index absolute error is less than 3 μ s, and relative error is less than 6%.

It should be noted that the listed above specific embodiment being only the present invention, it is clear that the present invention It is not limited to above example, has the similar change of many therewith.Those skilled in the art is such as All deformation that fruit is directly derived from present disclosure or associates, all should belong to this Bright protection domain.

Claims (4)

1. a replay method for transmission line malfunction current traveling wave waveform based on ripple weight, It is characterized in that, the method comprises the following steps:

Step S1: set multiple fault detecting point on transmission line of electricity successively, for fault electricity The collection of popular waveform；

Step S2: distance fault on transmission line of electricity is actually occurred place's homonymy except a nearest inspection The waveform of the another two fault detecting point of measuring point carries out card human relations boolean and converts three-phase decoupling；

Step S3: carry out after each modulus of row ripple after three-phase decouples is carried out wavelet package transforms Fast Fourier transform, point frequency band substitutes into circuit frequency according to function model；

Step S4: each frequency band is passed through the wave type energy of the detection waveform near trouble point Accounting is weighted, as coefficient, the target optimizing function that combination producing is final；

Step S5: by simulated annealing, target optimizing function is carried out the overall situation and solve, Distributed constant to this section of transmission line of electricity；

Step S6: the distributed constant that obtains is substituted into circuit frequency according to function model, according to known Point waveform combined circuit frequency is finally inversed by unknown point waveform according to parameter model.

Transmission line malfunction current traveling wave based on ripple weight the most according to claim 1 The replay method of waveform, it is characterised in that in described step S3, point frequency band substitutes into circuit Frequency is according to the concrete steps of function model:

Step S3.1: for after N shell wavelet package transforms, two test points are respectively divided into 2^NIndividual frequency Each frequency range waveform in Duan, carries out fast Fourier transform and generates respective 2^NSection frequency domain number According to.

Step S3.2: will be close to the frequency domain data of each modulus of test point at that of trouble point and be multiplied by Circuit according to frequency function H, and the data of the corresponding frequency band of modulus corresponding with another test point do poor Δ d (i), i represents i-th frequency range；

Wherein H=e^-λx, x is the distance between two test points, For each frequency range, f is the mid frequency of this frequency range, and R is resistance, and L is inductance, and G is electricity Leading, C is electric capacity.

Transmission line malfunction current traveling wave based on ripple weight the most according to claim 1 The replay method of waveform, it is characterised in that the concrete step of weighted array in described step S4 Rapid:

Step S4.1: for N shell wavelet package transforms, to the test point pointed out near fault Decoupling waveform carries out card human relations boolean and converts decoupling.

Step S4.2: each modulus waveform after decoupling is carried out N shell wavelet package transforms and is divided into 2^N Individual frequency range.

Step S4.3: the wave type energy that the component of each frequency range is had accounts for the hundred of total wave type energy Proportion by subtraction is exactly weight a (i) of weighted array, and i represents i-th frequency range；

Step S4.4: the object function after final weighting is as follows:* Represent conjugate transpose.

Transmission line malfunction current traveling wave ripple based on ripple weight the most according to claim 1 The method of shape, it is characterised in that the concrete step of inverting unknown point waveform in described step S6 Rapid:

S6.1: after obtaining circuit distributed constant in S5, determines that circuit frequency is according to function model H, wherein H=e^-λx, x is the distance of traveling wave, is entered by the known point waveform being used for inverting Row card human relations boolean converts decoupling becomes each modulus waveform；

S6.2: the traveling wave line mould waveform after decoupling three-phase carries out fast Fourier transform, according to The position relationship of unknown point and known point and circuit frequency do computing according to function model:

If unknown point is in the downstream of known point, then known point each modulus waveform is multiplied by circuit frequency according to letter Digital-to-analogue type H；

If unknown point is in the middle of known point and line fault point, then known point each modulus waveform divided by Circuit frequency is according to function model H；

S6.3: according to the line mould waveform of unknown point after computing, in conjunction with fault type boundary condition, Push away to obtain zero mould waveform of unknown point；

S6.4: carry out card human relations boolean's inverse transformation according to each modulus waveform of unknown point and obtain unknown point Three-phase fault current traveling wave waveform.