CN102721889B - Based on the cable incipient fault detection method of Phase information Singularity Detection - Google Patents

Abstract

The invention discloses a kind of cable incipient fault detection method based on Phase information Singularity Detection, its step is mainly: the current signal gathering cable to be detected in power distribution network, utilize Phase information to carry out N layer scattering wavelet transformation to current signal, obtain the wavelet conversion coefficient on each Decomposition order; Extract the amplitude information of the wavelet coefficient on each Decomposition order, follow the tracks of the modulus maximum of the wavelet coefficient amplitude on each Decomposition order, record the point that modulus maximum all appears in each Decomposition order: the correspondence of first modulus maximum point of record goes out to be the generation moment of initial failure now, it is then the finish time of initial failure that the correspondence of last modulus maximum point of record goes out now.The inventive method effectively can detect the initial failure of cable, to carry out early warning to defective cable, and take measures before fault occurs, greatly reduce the loss that permanent fault causes, improve the security of operation of power networks, extend cable operation life.

Description

Based on the cable incipient fault detection method of Phase information Singularity Detection

Technical field

The present invention relates to the method for a kind of cable incipient fault detection based on Phase information Singularity Detection in electric system.

Background technology

Along with the development of modern urban construction, the deal of cable shared by the net power supply of city is also more and more heavier, progressively instead of built on stilts power supply bare wire, is widely used in power distribution network.

In actual moving process, many consumers require harsher to power supply reliability, and electric power system breaks down and can cause very serious consequence; This just requires electric power system to break down to give warning in advance and can fast quick-recovery.Cable easily causes self permanent fault because of shelf depreciation, flashover etc. thus causes the fault of electric power system, when cable is about to permanent fault occurs, electric power system has been in dangerous state, now operator is difficult to electric power system operating mode to return to normal condition, namely enable operating mode is returned to normal condition, also can have a strong impact on the operation of electric power system.In fact, the cable in electric power system has a sizable part to be in insulation harm, local ageing, the junction contacts running status in spite of illness such as bad.Detect that the initial failure (band diseased state) of cable can carry out early warning to defective cable, and can take measures before fault occurs, the cable that there is hidden danger is avoided to repeat voltage transient, greatly reduce the loss that permanent fault causes, to the security, the prolongation cable operation life that improve operation of power networks, tool is of great significance.

Summary of the invention

The object of the invention is to propose a kind of cable incipient fault detection method based on Phase information Singularity Detection, the method effectively can detect the initial failure of cable, to carry out early warning to defective cable, and take measures before fault occurs, the cable that there is hidden danger is avoided to repeat voltage transient, greatly reduce the loss that permanent fault causes, to the security, the prolongation cable operation life that improve operation of power networks.

The present invention is for solving its technical matters, and the technical scheme adopted is a kind of cable incipient fault detection method based on Phase information Singularity Detection, the steps include:

A, gather the current signal of cable to be detected in power distribution network, more than sample frequency 10KHz, obtain sampled current signals i (t) of cable to be detected, wherein t represents sampling instant;

B, utilize Phase information to carry out N layer scattering wavelet decomposition to sampled current signals i (t), obtain the wavelet conversion coefficient W on the n-th decomposition layer _n(a, b), wherein, a is the scale parameter of wavelet transformation, and b is time displacement parameter, n=1,2 ..., N; N>=32;

Wavelet coefficient W on each decomposition layer of C, extract real-time _nthe amplitude of (a, b), and follow the tracks of the modulus maximum of the wavelet coefficient amplitude on each decomposition layer go out now; When each decomposition layer all occurs modulus maximum at synchronization, then judge that this moment there occurs initial failure, at least also have moment each decomposition layer all to occur modulus maximum subsequently, last each decomposition layer all occurs that the moment of modulus maximum is the finish time of initial failure.

Compared with prior art, the invention has the beneficial effects as follows:

When cable is in the band diseased states such as insulation harm, local ageing, junction contacts be bad, can there is the transient state jumping phenomenon of electric current and show initial failure at random.Initial failure stage of development, current signal be interrupted or certain order derivative discontinuous, the present invention finds out the moment that modulus maximum all appears in each decomposition layer effectively, accurately by complex wavelet transform, this moment is current signal and is interrupted or certain order derivative discontinuous moment, thus accurately finds generation and the finish time of cable initial failure.And then the order of severity of initial failure can be judged according to the occurrence frequency of initial failure and duration, provide corresponding early warning signal, operating personnel are and guided to take the measures such as corresponding reinforced insulation, reinforcing joint, voltage transient is repeated to avoid the cable that there is hidden danger, greatly reduce the generation of permanent fault, thus reduce the loss, improve the security of operation of power networks, extend cable operation life.

In above-mentioned B step, utilize Phase information to carry out N layer scattering wavelet decomposition to current signal i (t), obtain the wavelet conversion coefficient W on the n-th decomposition layer _nthe method of (a, b) is:

$W_n(a,b)={\∫}_{-\∞}^{+\∞}i\left(t\right)\·\frac{1}{\sqrt{a}}\stackrel{\‾}{\mathrm{\ψ}}\left(\frac{t-b}{a}\right)\mathrm{dt}$

In formula, a is the scale parameter of wavelet transformation, and b is time displacement parameter, and Ψ (t) is Phase information function, for the conjugation of Ψ (t).

In above-mentioned C step, extract the wavelet coefficient W on each Decomposition order _nthe amplitude of (a, b), the specific practice recording the modulus maximum of the wavelet coefficient amplitude on each Decomposition order is:

(1) the wavelet coefficient W on each decomposition layer is extracted _nthe amplitude of (a, b)

$A_{W_n}(n,t)=\sqrt{W_n{R}^2+W_n{I}^2}$

W in formula _nr and W _ni is respectively wavelet conversion coefficient W _nthe real part of (a, b) and imaginary part,

(2) the wavelet coefficient amplitude on each decomposition layer is recorded modulus maximum

$M_{W_n}(n,t^{\′})=\mathrm{Max}\left(A_{W_n}(n,t)\right)$

Wherein Max is the function of maximizing, and t ' represents the moment that the wavelet coefficient modulus maximum on Decomposition order n occurs.

Accompanying drawing explanation

The distribution test model topological structure of the emulation experiment of Fig. 1 embodiment of the present invention.

The structural representation of the cable feeder line of the emulation experiment of Fig. 2 embodiment of the present invention.

Sampled current signals in the half cycles emulation experiment of Fig. 3 a embodiment of the present invention.

The modulus maximum point distribution plan detected in the half cycles emulation experiment of Fig. 3 b embodiment of the present invention.

Sampled current signals in many cycles emulation experiment of Fig. 4 a embodiment of the present invention.

The modulus maximum point distribution plan detected in many cycles emulation experiment of Fig. 4 b embodiment of the present invention.

Embodiment

Below in conjunction with embodiment, the present invention is described in further detail.

Embodiment

The specific embodiment of the present invention is, a kind of cable incipient fault detection method based on Phase information Singularity Detection, the steps include:

A, gather the current signal of cable to be detected in power distribution network, more than sample frequency 10KHz, obtain sampled current signals i (t) of cable to be detected, wherein t represents sampling instant;

B, utilize Phase information to carry out N layer scattering wavelet decomposition to sampled current signals i (t), obtain the wavelet conversion coefficient W on the n-th decomposition layer _n(a, b), wherein, a is the scale parameter of wavelet transformation, and b is time displacement parameter, n=1,2 ..., N; N>=32:

Utilize Phase information to carry out N layer scattering wavelet decomposition to current signal i (t), obtain the wavelet conversion coefficient W on the n-th decomposition layer _nthe method of (a, b) is:

$W_n(a,b)={\∫}_{-\∞}^{+\∞}i\left(t\right)\·\frac{1}{\sqrt{a}}\stackrel{\‾}{\mathrm{\ψ}}\left(\frac{t-b}{a}\right)\mathrm{dt}$

In formula, a is the scale parameter of wavelet transformation, and b is time displacement parameter, and Ψ (t) is Phase information function, for the conjugation of Ψ (t).

Wavelet coefficient W on each decomposition layer of C, extract real-time _nthe amplitude of (a, b), and follow the tracks of the modulus maximum of the wavelet coefficient amplitude on each decomposition layer go out now; When each decomposition layer all occurs modulus maximum at synchronization, then judge that this moment there occurs initial failure, at least also have moment each decomposition layer all to occur modulus maximum subsequently, last each decomposition layer all occurs that the moment of modulus maximum is the finish time of initial failure.

Extract the wavelet coefficient W on each Decomposition order _nthe amplitude of (a, b), the specific practice recording the modulus maximum of the wavelet coefficient amplitude on each Decomposition order is:

(1) the wavelet coefficient W on each decomposition layer is extracted _nthe amplitude of (a, b)

$A_{W_n}(n,t)=\sqrt{W_n{R}^2+W_n{I}^2}$

W in formula _nr and W _ni is respectively wavelet conversion coefficient W _nthe real part of (a, b) and imaginary part,

(2) the wavelet coefficient amplitude on each decomposition layer is recorded modulus maximum

$M_{W_n}(n,t^{\′})=\mathrm{Max}\left(A_{W_n}(n,t)\right)$

Wherein Max is the function of maximizing, and t ' represents the moment that the wavelet coefficient modulus maximum on Decomposition order n occurs.

Emulation experiment:

For verifying the validity of the inventive method, carry out following emulation experiment.

Utilize the BergeronModel cable model in PSCAD/EMTDC software, the simplified model of distribution networks set up as shown in Figure 1 carries out modeling and simulating to cable initial failure, power supply is 110kV, reduce to 10kV connection through transformer and join 4 feeder lines, and be cable line, article 4, feeder line overall length 50km, distance bus 25km place arranges the simulating signal that switch 1 accesses initial failure.As shown in Figure 2, parameter is in table 1 for construction of cable schematic diagram.

Table 1 cable data

In emulation, sample frequency is 10kHz, selects bandwidth parameter to be 1.5, centre frequency be 1 multiple Morlet small echo the wavelet decomposition of N=32 layer is carried out to the current signal gathered.

Adopt above condition, carry out the emulation experiment of the detection of the detection of half cycles initial failure (initial failure of duration in power current half period) and many cycles initial failure initial failure of power current multiple cycle (duration be) respectively.

One, the emulation experiment of half cycles incipient fault detection:

Switch accesses the simulating signal of a half cycles initial failure.Fig. 3 a gives the sampled current signals that the inventive method obtains.Fig. 3 b gives the maximum value coordinate points of half cycles initial failure current signal wavelet coefficient amplitude on 1 ~ 32 decomposition layer, from two figure, the fault generation moment point detected is 1325 points, the corresponding time is 0.0825s, the moment point that fault terminates is 1384 points, the corresponding time is 0.0884s, and fault continue for 0.0059s.The fault start/stop time detected is consistent with the fault start/stop time in the half cycles initial failure signal of access.

Add to the half cycles initial failure current signal in above experiment the white Gaussian noise that signal to noise ratio (S/N ratio) is 70dB, 60dB, 50dB respectively, then detect three current signals after interpolation noise by the method for the present embodiment, testing result is as table 2.As can be seen from Table 2, when signal to noise ratio (S/N ratio) 70dB, 60dB and 50dB, detected result and the testing result of noise-free signal basically identical, noise does not cause too large impact to testing result.Illustrate that the noise immunity of the inventive method is good, reliability is high.

Initial failure positioning result under the different noise of table 2

Two, the emulation experiment of many cycles incipient fault detection:

The simulating signal of switch access cycle more than initial failure.Fig. 4 a gives the sampled current signals that the inventive method obtains, Fig. 4 b gives the maximum value coordinate points of wavelet coefficient amplitude on many cycles initial failure current signal 1 ~ 32 Decomposition order, can be read by two figure, the moment point that fault occurs is 1324 points, the corresponding time is 0.1324s, the moment point that fault terminates is 2104 points, and the corresponding time is 0.2104s, and fault continue for 0.078s.The fault start/stop time detected is consistent with the fault start/stop time in many cycles initial failure signal of access.

More than test proof, the inventive method can detect initial failure and the start/stop time thereof of cable accurately and efficiently.

Desirable more than 32 arbitrary integers of Decomposition order N in the present invention, value is larger, and Decomposition Accuracy is higher, but calculates more complicated.The optional frequency of desirable more than the 10kHz of sample frequency, sample frequency is higher, and the moment precision oriented is higher, and same calculated amount is also larger.

Claims (3)

1., based on a cable incipient fault detection method for Phase information Singularity Detection, the steps include:

A, gather the current signal of cable to be detected in power distribution network, more than sample frequency 10KHz, obtain sampled current signals i (t) of cable to be detected, wherein t represents sampling instant;

B, utilize Phase information to carry out N layer scattering wavelet decomposition to sampled current signals i (t), obtain the wavelet conversion coefficient W on the n-th decomposition layer _n(a, b), wherein, a is the scale parameter of wavelet transformation, and b is time displacement parameter, n=1,2 ..., N; N>=32;

Wavelet coefficient W on each decomposition layer of C, extract real-time _nthe amplitude of (a, b), and follow the tracks of the modulus maximum of the wavelet coefficient amplitude on each decomposition layer go out now; When each decomposition layer all occurs modulus maximum at synchronization, then judge that this moment there occurs initial failure, at least also have moment each decomposition layer all to occur modulus maximum subsequently, last each decomposition layer all occurs that the moment of modulus maximum is the finish time of initial failure.

2. the cable incipient fault detection method based on Phase information Singularity Detection according to claim 1, it is characterized in that, in described B step, utilize Phase information to carry out N layer scattering wavelet decomposition to sampled current signals i (t), obtain the wavelet conversion coefficient W on the n-th decomposition layer _nthe method of (a, b) is:

$W_n(a,b)={\∫}_{-\∞}^{+\∞}i\left(t\right)\·\frac{1}{\sqrt{\mathrm{\α}}}\stackrel{\‾}{\mathrm{\ψ}}\left(\frac{t-b}{a}\right)\mathrm{dt}$

In formula, a is the scale parameter of wavelet transformation, and b is time displacement parameter, and ψ (t) is Phase information function, for the conjugation of ψ (t).

3. the cable incipient fault detection method based on Phase information Singularity Detection according to claim 1, is characterized in that, in described C step, extracts the wavelet coefficient W on each Decomposition order _nthe amplitude of (a, b), the specific practice recording the modulus maximum of the wavelet coefficient amplitude on each Decomposition order is:

(1) the wavelet coefficient W on each decomposition layer is extracted _nthe amplitude of (a, b)

$A_{W_n}(n,t)=\sqrt{W_n{R}^2+W_n{I}^2}$

W in formula _nr and W _ni is respectively wavelet conversion coefficient W _nthe real part of (a, b) and imaginary part,

(2) the wavelet coefficient amplitude on each decomposition layer is recorded modulus maximum

$M_{W_n}(n,t^{\′})=\mathrm{Max}\left(A_{W_n}(x,t)\right)$

Wherein Max is the function of maximizing, and t ' represents the moment that the modulus maximum of the wavelet coefficient amplitude on Decomposition order n occurs.