CN104122486A - Method and device for detecting early failure of cable - Google Patents

Abstract

The invention provides a method and a device for detecting the early failure of a cable. The method comprises the steps of orderly performing wavelet analysis on a sampling window comprising a preset number of sampling points, and if the criterion of a first transient process and the criterion of a second transient process are satisfied after the wavelet analysis and the time interval of the two transient processes is within 1/4 cyclic wave and 5 cyclic waves, determining the early failure between the first transient process and the second transient process and considering the first occurrence moment of the first transient process as the starting time of the early failure and the second occurrence moment of the first transient process as the terminal time of the early failure. As a result, the method is capable of accurately identifying the early failure and giving the exact starting and terminal moments of the failure.

Description

A kind of cable incipient fault detection method and device

Technical field

The present invention relates to power cable field, relate in particular to a kind of cable incipient fault detection method and device.

Background technology

Along with the continuous propelling of Process of Urbanization Construction, urban power distribution network trends towards adopting power cable to carry out electric power transfer day by day at present.The fault progression process of cable, can be divided into initial failure and permanent fault.Initial failure is normally inner aging and run down and cause by insulation.If initial failure can not obtain in time detection and Identification and be repaired, will develop into permanent fault.

Power cable initial failure is mainly by the time domain of voltage and current amount, the feature detection of frequency domain at present.Temporal signatures detection method mainly changes by the amplitude of detection streamer voltage and current, determine whether as initial failure, the method can provide the beginning and ending time of each fault, yet the feature of initial failure is mainly reflected in the high fdrequency component of voltage and current signal, only from voltage and current amplitude, changes and be difficult to accurately follow the tracks of initial failure.Whether frequency domain character detection method is mainly differentiated by the each harmonic component of detection voltage and current is initial failure, but can not clearly give the start-stop moment of being out of order and occurring.

Therefore need now a kind of detection method, can either accurately differentiate initial failure again can be clearly to out of order start-stop constantly.

Summary of the invention

The invention provides a kind of detection method, Apparatus and system of cable initial failure, this method can accurately be differentiated initial failure again can be clearly to out of order start-stop constantly.

To achieve these goals, the invention provides following technological means:

A cable incipient fault detection method, comprising:

By predeterminated frequency, the current signal of cable is sampled, obtain several sampled points in a cycle;

Utilize wavelet transform to carry out three yardstick Mallat to a sample window and decompose, obtain high frequency detail coefficients and low frequency Coefficients of Approximation under three yardsticks, described sample window comprises the sampled point of predetermined number;

After above-mentioned wavelet analysis, if one of three criterions of the satisfied transient state process for the first time of described high frequency detail coefficients or described low frequency Coefficients of Approximation, be defined as transient state process for the first time, and recording the first generation constantly, described three criterions comprise that the reconstruction signal that root-mean-square value that described high frequency detail coefficients energy value is greater than the first preset value, described low frequency Coefficients of Approximation is greater than the second preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

Described sample window is rolled to next sampled point, extract the sampled point of predetermined number as current window, window corresponding to transient state process is called history window for the first time, and described current window is carried out to wavelet analysis equally, obtains current high frequency detail coefficients and current low frequency Coefficients of Approximation;

If the time interval of described current window and history window is within 1/4 cycle～5 cycle, described current high frequency detail coefficients or described current low frequency Coefficients of Approximation meet one of three criterions of transient state process for the second time, confirm to occur transient state process for the second time, and recording the second generation constantly, three criterions of described transient state process for the second time comprise that the reconstruction signal that root-mean-square value that described current high frequency detail coefficients energy value is greater than the 3rd preset value, described low frequency Coefficients of Approximation is greater than the 3rd preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

Described transient state process for the first time and described between transient state process, be initial failure for the second time, described first to occur constantly be initial failure initial time, described second the moment occurs is the initial failure termination time.

Preferably, also comprise:

If the time interval of described current window and history window within 1/4 cycle or while being greater than 5 cycles, judges described current high frequency detail coefficients or described current low frequency Coefficients of Approximation and whether meets one of three criterions of transient state process for the first time.

Preferably, described the 3rd preset value is 70% of described the first preset value, and described the 4th preset value is 70% of described the second preset value.

Preferably, described predeterminated frequency comprises 3.2kHz, at predeterminated frequency, be under 3.2kHz, frequency range corresponding to 3 yardstick high frequency detail coefficients is respectively 0.8～1.6kHz, 0.4～0.8kHz and 0.2～0.4kHz, and frequency range corresponding to the 3rd yardstick low frequency Coefficients of Approximation is 0～0.2kHz.

Preferably, described high frequency detail coefficients energy value is greater than the first preset value and comprises:

$E_d=\frac{E_{d,l}-{\stackrel{\‾}{E}}_{d,1~l-1}}{\mathrm{\σ}\left(E_{d,1~l-1}\right)}>F_{M1}$

Its, E _d,lthe high frequency detail coefficients energy value sum recording for current window, the average of the detail coefficients energy value sum recording for history window, σ (E _{d, 1～l-1}) standard deviation of the high frequency detail coefficients energy value sum that records for history window, F _m1for described the first preset value.

Preferably, described low frequency Coefficients of Approximation root-mean-square value is greater than the second preset value and comprises:

$R_c=\left|\frac{R_{c,l}-R_{c,l-64}}{R_{c,l-64}}\right|>F_{M2}$

Wherein, R _c,lrepresent 0～0.2kHz frequency band that current window records, the root-mean-square value of low frequency Coefficients of Approximation, R _{c, l-64}the root-mean-square value of the 0～0.2kHz frequency band Coefficients of Approximation recording for the corresponding window of a upper cycle, F _m2for described the second preset value.

Preferably, the reconstruction signal that the 3rd yardstick approaches exists modulus maximum to comprise:

Second order after signal after reconstruct and wavelet function convolution is led and is had zero crossing.

Preferably, each sample window comprises 8 sampled points, and each sample window rolls and reads current sampling point and current sampling point 7 sampled points before.

Preferably, also comprise:

Occur after initial failure, counter adds one, and carries out the detection of initial failure next time.

A cable incipient fault detection device, comprising:

The first wavelet analysis unit, for by predeterminated frequency, the current signal of cable being sampled, obtains several sampled points in a cycle; Utilize wavelet transform to carry out three yardstick Mallat to a sample window and decompose, obtain high frequency detail coefficients and low frequency Coefficients of Approximation under three yardsticks, described sample window comprises the sampled point of predetermined number;

The first judging unit, for after above-mentioned wavelet analysis, if one of three criterions of the satisfied transient state process for the first time of described high frequency detail coefficients or described low frequency Coefficients of Approximation, be defined as transient state process for the first time, and recording the first generation constantly, described three criterions comprise that the reconstruction signal that root-mean-square value that described high frequency detail coefficients energy value is greater than the first preset value, described low frequency Coefficients of Approximation is greater than the second preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

The second wavelet analysis unit, for described sample window is rolled to next sampled point, extract the sampled point of predetermined number as current window, window corresponding to transient state process is called history window for the first time, described current window is carried out to wavelet analysis equally, obtain current high frequency detail coefficients and current low frequency Coefficients of Approximation;

The second judging unit, if be used for the time interval of described current window and history window within 1/4 cycle～5 cycle, described current high frequency detail coefficients or described current low frequency Coefficients of Approximation meet one of three criterions of transient state process for the second time, confirm to occur transient state process for the second time, and record the second generation constantly, three criterions of described transient state process for the second time comprise that described current high frequency detail coefficients energy value is greater than the 3rd preset value, there is modulus maximum in the reconstruction signal that the root-mean-square value of described low frequency Coefficients of Approximation is greater than the 3rd preset value and the 3rd yardstick Coefficients of Approximation,

Record cell, for described transient state process for the first time and described between transient state process, be initial failure for the second time, described first to occur constantly be initial failure initial time, described second the moment occurs is the initial failure termination time.

The invention provides a kind of detection method and device of initial failure, the present invention samples to current signal, successively to comprising that the sampled point sample window of predetermined number carries out wavelet analysis, if meet the criterion of transient state process and the criterion of the second transient state process for the first time after wavelet analysis, and the time interval of twice transient state process is within 1/4 cycle～5 cycle, between definite transient state process for the first time and the second transient state process, be initial failure, and, first of transient state process the generation is initial failure initial time constantly for the first time, second of transient state process the generation is the initial failure termination time constantly for the second time.Therefore the present invention can accurately differentiate initial failure again can be clearly to out of order start-stop constantly.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, to the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skills, do not paying under the prerequisite of creative work, can also obtain according to these accompanying drawings other accompanying drawing.

Fig. 1 is the detection method process flow diagram of the disclosed a kind of initial failure of the embodiment of the present invention;

Fig. 2 is the schematic diagram of sample window in the detection method of the disclosed a kind of initial failure of the embodiment of the present invention;

Fig. 3 is the schematic diagram of wavelet analysis method in the detection method of the disclosed a kind of initial failure of the embodiment of the present invention;

Fig. 4 is the process flow diagram of the detection method of disclosed another initial failure of the embodiment of the present invention;

Fig. 5 is the structural representation of the pick-up unit of the disclosed a kind of initial failure of the embodiment of the present invention.

Embodiment

Before the embodiment of the present invention, paper wavelet analysis:

Wavelet analysis is that a kind of window area is fixed, the changeable Time-Frequency Localization analytical approach of shape.In low frequency part, there is higher frequency resolution and lower temporal resolution, at HFS, there is higher temporal resolution and lower frequency resolution.A given basic function ψ (t), wavelet function is through translation and wavelet function collection flexible and that form by female small echo ψ (t).Order

${\mathrm{\ψ}}_{a,b}\left(t\right)=\frac{1}{\sqrt{a}}\mathrm{\ψ}\left(\frac{t-b}{a}\right)---\left(1\right)$

A in formula (1), b is constant, and a>0.Obviously, ψ _a,b(t) be that basic function ψ (t) first does displacement and remakes after flexible and obtain.If a, b constantly changes, and we can obtain the function ψ of gang _a,b(t).

Given square-integrable signal x (t), i.e. x (t) ∈ L ²(R), the wavelet transformation of x (t) (Wavelet Transform, WT) is defined as

$\begin{array}{c}{\mathrm{WT}}_x(a,b)=\frac{1}{\sqrt{a}}{\&Integral;}_{-\&infin;}^{+\&infin;}x\left(t\right){\mathrm{\&psi;}}^{*}\left(\frac{t-b}{a}\right)\mathrm{dt}\\ ={\&Integral;}_{-\&infin;}^{+\&infin;}x\left(t\right){\mathrm{\&psi;}}_{a,b}^{*}\left(t\right)\mathrm{dt}=<x\left(t\right),{\mathrm{\&psi;}}_{a,b}\left(t\right)>\end{array}---\left(2\right)$

A in formula, b and t are all continuous variables, so this formula is called again continuous wavelet transform (continuous wavelet transform, CWT).The wavelet transformation WT of signal x (t) _x(a, b) is the function of a and b, and a is scale factor, and b is time shift.ψ (t) is called again wavelet or female small echo.ψ _a,b(t) be the gang function of female small echo through being shifted and stretching and produce, be referred to as wavelet basis function, or be called for short wavelet basis, for ψ _a,b(t) complex conjugate.

Scale factor a is entered to mode discretize by two, time shift b is pressed to binary integer mode discretize doubly, i.e. a=2 ^j, b=n2 ^j, the wavelet basis function progression that can obtain quadrature is

ψ _j,n(t)＝2 ^-j/2ψ(2 ^-jt-n) (3)

The wavelet transform of discrete signal x (i) (discrete wavelet transform, DWT) can be expressed as so

${\mathrm{DWT}}_x[j,n]=\frac{1}{\sqrt{2^j}}\underset{k=-\&infin;}{\overset{+\&infin;}{\mathrm{\&Sigma;}}}x\left[i\right]\mathrm{\&psi;}\left(\frac{k-{n2}^j}{2^j}\right)---\left(4\right)$

Wherein, i is positive integer, is signal sampling point sequence number.

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is only the present invention's part embodiment, rather than whole embodiment.Embodiment based in the present invention, those of ordinary skills, not making the every other embodiment obtaining under creative work prerequisite, belong to the scope of protection of the invention.

As shown in Figure 1, the invention provides a kind of cable incipient fault detection method, comprising:

Step S101: by predeterminated frequency, the current signal of cable is sampled, obtain several sampled points in a cycle;

Step S102: utilize wavelet transform to carry out three yardstick Mallat to a sample window and decompose, obtain high frequency detail coefficients and low frequency Coefficients of Approximation under three yardsticks, described sample window comprises the sampled point of predetermined number;

As shown in Figure 2, be the schematic diagram of sample window, in Fig. 2, take sample window as 8 sampled points be example, a window is l, next window is l+1, a sampled point is i, next sampled point is i+1; Utilizing wavelet transform to carry out three yardstick Mallat to a sample window decomposes to be and carries out wavelet analysis one time.

Step S103: after above-mentioned wavelet analysis, if one of three criterions of the satisfied transient state process for the first time of described high frequency detail coefficients or described low frequency Coefficients of Approximation are defined as transient state process for the first time, and record the first generation constantly.

Wherein, described three criterions comprise that the reconstruction signal that root-mean-square value that described high frequency detail coefficients energy value is greater than the first preset value, described low frequency Coefficients of Approximation is greater than the second preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

Step S104: described sample window is rolled to next sampled point, extract the sampled point of predetermined number as current window, window corresponding to transient state process is called history window for the first time, described current window is carried out to wavelet analysis equally, obtain current high frequency detail coefficients and current low frequency Coefficients of Approximation;

Step S105: the time interval of judgement current window and history window is within 1/4 cycle to 5 cycle; If enter step S106, otherwise enter step S103.

Preferably, if the time interval of described current window and history window within 1/4 cycle or while being greater than 5 cycles, judges described current high frequency detail coefficients or described current low frequency Coefficients of Approximation and whether meets one of three criterions of transient state process for the first time.

Step S106: judge one of three criterions of the satisfied transient state process for the second time of described current high frequency detail coefficients or described current low frequency Coefficients of Approximation; If enter step S107, otherwise enter step S103;

Step S107: confirm transient state process for the second time occurs, and record the second generation constantly;

If the time interval of described current window and history window is within 1/4 cycle～5 cycle, described current high frequency detail coefficients or described current low frequency Coefficients of Approximation meet one of three criterions of transient state process for the second time, described in for the second time three criterions of transient state process comprise that the reconstruction signal that root-mean-square value that described current high frequency detail coefficients energy value is greater than the 3rd preset value, described low frequency Coefficients of Approximation is greater than the 3rd preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

Step S108: described transient state process for the first time and described between transient state process, be initial failure for the second time, described first to occur constantly be initial failure initial time, described second the moment occurs is the initial failure termination time.

Preferably, described the 3rd preset value is 70% of described the first preset value, and described the 4th preset value is 70% of described the second preset value.

The invention provides a kind of detection method of initial failure, the present invention samples to current signal, successively to comprising that the sampled point sample window of predetermined number carries out wavelet analysis, if meet the criterion of transient state process and the criterion of the second transient state process for the first time after wavelet analysis, and the time interval of twice transient state process is within 1/4 cycle～5 cycle, between definite transient state process for the first time and the second transient state process, be initial failure, and, first of transient state process the generation is initial failure initial time constantly for the first time, second of transient state process the generation is the initial failure termination time constantly for the second time.Therefore the present invention can accurately differentiate initial failure again can be clearly to out of order start-stop constantly.

Introduce a specific embodiment of the present invention below:

Step S201: current signal is arranged to the sample frequency of 3.2kHz, every cycle is adopted 64 points;

Step S202: each window is got continuous 8 sampled points, the sequence number of note window end sampled point is i;

The present invention samples to the current signal of power cable, and the proportion of setting is 3.2kHz, and every cycle is adopted 64 points.With reference to accompanying drawing 2, sample window rolls and reads current sampling point i, and continuous 7 points before current sampling point i, then carries out discrete wavelet analysis.

Step S203: make i=1;

Step S204: this window is designated as t constantly _w, select discrete wavelet variation to carry out three layers of Mallat decomposition to 8 sampled points of this window and obtain each high band detail coefficients and low-frequency range approximation coefficient;

Step S205: judge whether 8 sampled points of this time meet the criterion of transient state process for the first time; If enter step S207, otherwise enter step S206;

8 sampled points that current window is read carry out Mallat decomposition, through three yardstick resolution decomposition operation, signal decomposition are become to low frequency Coefficients of Approximation and high frequency detail coefficients.The detail coefficients d of current signal under j decomposition scale _j,nwith Coefficients of Approximation c _j,nfor:

$\left\{\begin{array}{c}{c}_{j,n}=\underset{i}{\mathrm{\&Sigma;}}{h}_{i-2n}{c}_{j-1,i}\\ d_{j,n}=\underset{i}{\mathrm{\&Sigma;}}{g}_{i-2n}{c}_{j-1,i}\end{array}\right.,j=\mathrm{1,2,3}---\left(5\right)$

Wherein h and g are respectively low pass and Hi-pass filter, above formula shows, detail coefficients on j yardstick and Coefficients of Approximation can have the Coefficients of Approximation on j-1 yardstick to carry out convolution with high and low pass filter respectively to carry out after two extractions obtaining again, and its implementation procedure block diagram as shown in Figure 3.

Because sample frequency is 3.2kHz, the frequency range that therefore 1-3 yardstick detail coefficients and the 3rd yardstick Coefficients of Approximation are corresponding is as shown in table 1.

The frequency range that each scale coefficient of table 1 is corresponding

Each scale coefficient	d1	d2	d3	c3
					Frequency range (kHz)	0.8～1.6	0.4～0.8	0.2～0.4	0～0.2

Current window signal is carried out after Mallat decomposition, utilize the detail coefficients of its 1-3 yardstick and the Coefficients of Approximation of the 3rd yardstick to carry out transient state process for the first time and detect.Below meeting, during one of three criterions, be judged to be the generation of transient state process, and to record current time be t1.

Criterion one: if the detail coefficients of 0.2～1.6kHz frequency band meets following formula, testing result is transient state process.

$E_d=\frac{E_{d,l}-{\stackrel{\&OverBar;}{E}}_{d,1~l-1}}{\mathrm{\&sigma;}\left(E_{d,1~l-1}\right)}>F_{M1}---\left(6\right)$

E wherein _d,lthe detail coefficients energy value sum recording for current window, the average of the detail coefficients energy value sum recording for history window, σ (E _{d, 1～l-1}) standard deviation of the detail coefficients energy value sum that records for history window, F _m1threshold values for criterion one setting.

The detail coefficients d of j yardstick _j,nenergy value E _d,jcan be expressed as

$E_{d,j}=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{d}_{j,n}^2---\left(7\right)$

The energy value sum of detail coefficients under the 1-3 yardstick of window l so

$E_{d,l}=\underset{j=1}{\overset{3}{\mathrm{\&Sigma;}}}{E}_{d,j}=\underset{j=1}{\overset{3}{\mathrm{\&Sigma;}}}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{d}_{j,n}^2---\left(8\right)$

Criterion two: if the Coefficients of Approximation of 0～0.2kHz frequency band meets following formula, testing result is transient state process.

$R_c=\left|\frac{R_{c,l}-R_{c,l-64}}{R_{c,l-64}}\right|>F_{M2}---\left(9\right)$

R wherein _c,lthe root-mean-square value that represents 0～0.2kHz frequency band Coefficients of Approximation that current window records, R _{c, l-64}the root-mean-square value of the 0～0.2kHz frequency band Coefficients of Approximation recording for the corresponding window of a upper cycle, F _m2threshold values for criterion two settings.

$R_{c,l}=\sqrt{\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{c}_{3,n}^2}---\left(10\right)$

Criterion three: if the reconstruction signal of the 3rd yardstick exists modulus maximum, testing result is transient state process.Concrete discriminant approach is as follows:

Utilize the Coefficients of Approximation of the 3rd yardstick to carry out the reconstruct of current signal, by the reconstruction signal x'(t obtaining) with wavelet function ψ (t) convolution after and ask for its single order and lead with second order and lead, it is the wavelet modulus maxima point that 0 point is reconstruction signal that its second order is led.

Step S206: make i=i+1, enter step S204;

Step S207: record transient state process initial time t for the first time ₁;

Step S208:i=i+1, this window is designated as t constantly _w;

Step S209: judgement t _w-t ₁whether be less than 1/4 cycle; If enter step S210, otherwise enter step S204;

Step S210: judgement t _w-t ₁whether be greater than 5 cycles; If enter step S211; Otherwise enter step S204;

As current window moment t _wwith t ₁difference when being less than 1/4 cycle or being greater than 5 cycles, detail coefficients and Coefficients of Approximation continue on for transient state process for the first time and detect, if testing result meets transient state process for the first time, again record t ₁, repeat the testing process that said process carries out next window.

Step S211: select discrete wavelet variation to carry out three layers of Mallat decomposition to 8 sampled points of this window and obtain each high band detail coefficients and low-frequency range approximation coefficient;

Step S212: judge whether 8 sampled points of this time meet the criterion of transient state process for the second time; If satisfied enter step S213, otherwise enter step S208;

Then rolling window, repeats to read detail coefficients and the Coefficients of Approximation after window sampled signal wavelet transformation and Mallat decomposition, as current window moment t _wwith t ₁difference within 1/4～5 cycle time, detail coefficients and Coefficients of Approximation are for the detection of transient state process for the second time, discrimination standard is identical with transient state process for the first time, just the threshold values of the first two standard setting is primary 70%, if while meeting one of three criterions, be transient state process for the second time, and record moment t ₂;

Step S213: record transient state process initial time t for the second time ₂;

Step S214:t ₁to t ₂in period, be initial failure one time, counter records once, and is carried out fault detect next time.Enter step S204.

The invention provides a kind of detection method and device of initial failure, the present invention samples to current signal, successively to comprising that the sampled point sample window of predetermined number carries out wavelet analysis, if meet the criterion of transient state process and the criterion of the second transient state process for the first time after wavelet analysis, and the time interval of twice transient state process is within 1/4 cycle～5 cycle, between definite transient state process for the first time and the second transient state process, be initial failure, and, first of transient state process the generation is initial failure initial time constantly for the first time, second of transient state process the generation is the initial failure termination time constantly for the second time.Therefore the present invention can accurately differentiate initial failure again can be clearly to out of order start-stop constantly.

As shown in Figure 5, the present invention also provides a kind of cable incipient fault detection device, comprising:

The first wavelet analysis unit 100, for by predeterminated frequency, the current signal of cable being sampled, obtains several sampled points in a cycle; Utilize wavelet transform to carry out three yardstick Mallat to a sample window and decompose, obtain high frequency detail coefficients and low frequency Coefficients of Approximation under three yardsticks, described sample window comprises the sampled point of predetermined number;

The first judging unit 200, for after above-mentioned wavelet analysis, if one of three criterions of the satisfied transient state process for the first time of described high frequency detail coefficients or described low frequency Coefficients of Approximation, be defined as transient state process for the first time, and recording the first generation constantly, described three criterions comprise that the reconstruction signal that root-mean-square value that described high frequency detail coefficients energy value is greater than the first preset value, described low frequency Coefficients of Approximation is greater than the second preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

The second wavelet analysis unit 300, for described sample window is rolled to next sampled point, extract the sampled point of predetermined number as current window, window corresponding to transient state process is called history window for the first time, described current window is carried out to wavelet analysis equally, obtain current high frequency detail coefficients and current low frequency Coefficients of Approximation;

The second judging unit 400, if be used for the time interval of described current window and history window within 1/4 cycle～5 cycle, described current high frequency detail coefficients or described current low frequency Coefficients of Approximation meet one of three criterions of transient state process for the second time, confirm to occur transient state process for the second time, and record the second generation constantly, three criterions of described transient state process for the second time comprise that described current high frequency detail coefficients energy value is greater than the 3rd preset value, there is modulus maximum in the reconstruction signal that the root-mean-square value of described low frequency Coefficients of Approximation is greater than the 3rd preset value and the 3rd yardstick Coefficients of Approximation,

Record cell 500, for described transient state process for the first time and described between transient state process, be initial failure for the second time, described first to occur constantly be initial failure initial time, described second the moment occurs is the initial failure termination time.

If the function described in the present embodiment method usings that the form of SFU software functional unit realizes and during as production marketing independently or use, can be stored in a computing equipment read/write memory medium.Understanding based on such, the part that the embodiment of the present invention contributes to prior art or the part of this technical scheme can embody with the form of software product, this software product is stored in a storage medium, comprise that some instructions are with so that a computing equipment (can be personal computer, server, mobile computing device or the network equipment etc.) carry out all or part of step of method described in each embodiment of the present invention.And aforesaid storage medium comprises: USB flash disk, portable hard drive, ROM (read-only memory) (ROM, Read-Only Memory), the various media that can be program code stored such as random access memory (RAM, Random Access Memory), magnetic disc or CD.

In this instructions, each embodiment adopts the mode of going forward one by one to describe, and each embodiment stresses is the difference with other embodiment, between each embodiment same or similar part mutually referring to.

Above-mentioned explanation to the disclosed embodiments, makes professional and technical personnel in the field can realize or use the present invention.To the multiple modification of these embodiment, will be apparent for those skilled in the art, General Principle as defined herein can, in the situation that not departing from the spirit or scope of the present invention, realize in other embodiments.Therefore, the present invention will can not be restricted to these embodiment shown in this article, but will meet the widest scope consistent with principle disclosed herein and features of novelty.

Claims (10)

1. a cable incipient fault detection method, is characterized in that, comprising:

By predeterminated frequency, the current signal of cable is sampled, obtain several sampled points in a cycle;

Utilize wavelet transform to carry out three yardstick Mallat to a sample window and decompose, obtain high frequency detail coefficients and low frequency Coefficients of Approximation under three yardsticks, described sample window comprises the sampled point of predetermined number;

After above-mentioned wavelet analysis, if one of three criterions of the satisfied transient state process for the first time of described high frequency detail coefficients or described low frequency Coefficients of Approximation, be defined as transient state process for the first time, and recording the first generation constantly, described three criterions comprise that the reconstruction signal that root-mean-square value that described high frequency detail coefficients energy value is greater than the first preset value, described low frequency Coefficients of Approximation is greater than the second preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

Described sample window is rolled to next sampled point, extract the sampled point of predetermined number as current window, window corresponding to transient state process is called history window for the first time, and described current window is carried out to wavelet analysis equally, obtains current high frequency detail coefficients and current low frequency Coefficients of Approximation;

If the time interval of described current window and history window is within 1/4 cycle～5 cycle, described current high frequency detail coefficients or described current low frequency Coefficients of Approximation meet one of three criterions of transient state process for the second time, confirm to occur transient state process for the second time, and recording the second generation constantly, three criterions of described transient state process for the second time comprise that the reconstruction signal that root-mean-square value that described current high frequency detail coefficients energy value is greater than the 3rd preset value, described low frequency Coefficients of Approximation is greater than the 3rd preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

Described transient state process for the first time and described between transient state process, be initial failure for the second time, described first to occur constantly be initial failure initial time, described second the moment occurs is the initial failure termination time.

2. the method for claim 1, is characterized in that, also comprises:

If the time interval of described current window and history window within 1/4 cycle or while being greater than 5 cycles, judges described current high frequency detail coefficients or described current low frequency Coefficients of Approximation and whether meets one of three criterions of transient state process for the first time.

3. the method for claim 1, is characterized in that, described the 3rd preset value is 70% of described the first preset value, and described the 4th preset value is 70% of described the second preset value.

4. the method for claim 1, it is characterized in that, described predeterminated frequency comprises 3.2kHz, at predeterminated frequency, be under 3.2kHz, frequency range corresponding to 3 yardstick high frequency detail coefficients is respectively 0.8～1.6kHz, 0.4～0.8kHz and 0.2～0.4kHz, and frequency range corresponding to the 3rd yardstick low frequency Coefficients of Approximation is 0～0.2kHz.

5. the method for claim 1, is characterized in that, described high frequency detail coefficients energy value is greater than the first preset value and comprises:

$E_d=\frac{E_{d,l}-{\stackrel{\&OverBar;}{E}}_{d,1~l-1}}{\mathrm{\&sigma;}\left(E_{d,1~l-1}\right)}>F_{M1}$

Its, E _d,lthe high frequency detail coefficients energy value sum recording for current window, the average of the detail coefficients energy value sum recording for history window, σ (E _{d, 1～l-1}) standard deviation of the high frequency detail coefficients energy value sum that records for history window, F _m1for described the first preset value.

6. the method for claim 1, is characterized in that, described low frequency Coefficients of Approximation root-mean-square value is greater than the second preset value and comprises:

$R_c=\left|\frac{R_{c,l}-R_{c,l-64}}{R_{c,l-64}}\right|>F_{M2}$

Wherein, R _c,lrepresent 0～0.2kHz frequency band that current window records, the root-mean-square value of low frequency Coefficients of Approximation, R _{c, l-64}the root-mean-square value of the 0～0.2kHz frequency band Coefficients of Approximation recording for the corresponding window of a upper cycle, F _m2for described the second preset value.

7. the method for claim 1, is characterized in that, the reconstruction signal that the 3rd yardstick approaches exists modulus maximum to comprise:

Second order after signal after reconstruct and wavelet function convolution is led and is had zero crossing.

8. the method for claim 1, is characterized in that, each sample window comprises 8 sampled points, and each sample window rolls and reads current sampling point and current sampling point 7 sampled points before.

9. the method for claim 1, is characterized in that, also comprises:

Occur after initial failure, counter adds one, and carries out the detection of initial failure next time.

10. a cable incipient fault detection device, is characterized in that, comprising:

The first wavelet analysis unit, for by predeterminated frequency, the current signal of cable being sampled, obtains several sampled points in a cycle; Utilize wavelet transform to carry out three yardstick Mallat to a sample window and decompose, obtain high frequency detail coefficients and low frequency Coefficients of Approximation under three yardsticks, described sample window comprises the sampled point of predetermined number;

The first judging unit, for after above-mentioned wavelet analysis, if one of three criterions of the satisfied transient state process for the first time of described high frequency detail coefficients or described low frequency Coefficients of Approximation, be defined as transient state process for the first time, and recording the first generation constantly, described three criterions comprise that the reconstruction signal that root-mean-square value that described high frequency detail coefficients energy value is greater than the first preset value, described low frequency Coefficients of Approximation is greater than the second preset value and the 3rd yardstick Coefficients of Approximation exists modulus maximum;

The second wavelet analysis unit, for described sample window is rolled to next sampled point, extract the sampled point of predetermined number as current window, window corresponding to transient state process is called history window for the first time, described current window is carried out to wavelet analysis equally, obtain current high frequency detail coefficients and current low frequency Coefficients of Approximation;

The second judging unit, if be used for the time interval of described current window and history window within 1/4 cycle～5 cycle, described current high frequency detail coefficients or described current low frequency Coefficients of Approximation meet one of three criterions of transient state process for the second time, confirm to occur transient state process for the second time, and record the second generation constantly, three criterions of described transient state process for the second time comprise that described current high frequency detail coefficients energy value is greater than the 3rd preset value, there is modulus maximum in the reconstruction signal that the root-mean-square value of described low frequency Coefficients of Approximation is greater than the 3rd preset value and the 3rd yardstick Coefficients of Approximation,

Record cell, for described transient state process for the first time and described between transient state process, be initial failure for the second time, described first to occur constantly be initial failure initial time, described second the moment occurs is the initial failure termination time.