CN104198901A - Locating method and system for partial discharge signal of substation - Google Patents

Abstract

The invention discloses a locating method and a locating system for a partial discharge signal of a substation. The method comprises the steps of acquiring the partial discharge signal generated by the substation through at least four UHF (Ultrahigh Frequency) sensors; performing de-noising processing on the local discharge signal by using a preset complex wavelet function; using a preset high-order cumulant function to calculate the time delay of the local discharge signal in the different UHF sensors after the de-noising processing; calculating the generating position of the local discharge signal in accordance with the time delay and the installation positions of all UHF sensors. According to the locating method provided by the invention, the accuracy and reliability of locating the local discharge signal of the substation can be improved, and the problem that the UHF local discharge signal of an opening substation is complicate to locate in the traditional technology can be solved.

Description

The localization method of transformer station partial discharge signals and system

Technical field

The present invention relates to shelf depreciation technical field, particularly relate to a kind of localization method of transformer station partial discharge signals, and a kind of positioning system of transformer station partial discharge signals.

Background technology

Local discharge signal in electric power equipment detection method is mainly to adopt pulse current method, supersonic testing method, ultrahigh frequency detection method.

Highly sensitive and the sensor of pulse current method is easily installed, but poor anti jamming capability.When power equipment produces partial discharge phenomenon in transformer station, can follow various interference, additional undesired signal can exert an influence to detecting.

Power equipment can produce ultrasound wave while producing shelf depreciation, and Ultrasonic Detection ratio juris is ultrasound emission sensor to be installed on power equipment identify, locate shelf depreciation defect.Ultrasonic signal decay is more serious, and this detection method precision in the situation that sensor approaches shelf depreciation point is large, and away from accident defect in the situation that, the sensitivity of the method reduces greatly.

Ultrahigh frequency method is the new detection method of partial discharge of electrical equipment proposing in recent years, is developed rapidly.The pulse signal rise time that during due to generation shelf depreciation, in insulator, bubble electric discharge produces is very short, approximately 0.35-3ns, and pulsewidth 1-5ns, the electromagnetic pulse signal frequency band that therefore shelf depreciation produces should be more than hundreds of MHz.And the ground unrest frequency being produced by line corona, carrier communication, radio broadcasting etc. in transformer station is conventionally in several MHz left and right, partial discharge monitoring system bandwidth based on ultrahigh frequency method can reach hundreds of MHz to upper GHz, highly sensitive, can effectively avoid the impact of low frequency range noise, and can realize non-contact measurement, less on primary equipment impact.For open type transformer station, all may there is shelf depreciation in any high-tension apparatus in standing, if each equipment is installed to local discharge on-line monitoring device, cost is high, and causes the wasting of resources.

In conventional art, the collection of local signal can be installed four uhf sensors at automobile top, vehicle-mounted partial discharge detecting system goes on patrol near each equipment in open type transformer station, and the equipment that can effectively detect is airborne ultra-high frequency signal because shelf depreciation is radiated.Due to the difference of travel path, local discharge signal arrives the asynchronism(-nization) of each sensor, utilizes the mistiming forming between each sensor, can calculate the position of Partial Discharge Sources.This device can detect in movement, can realize the electrical equipment of whole open type transformer station is carried out to partial discharge monitoring, has that antijamming capability is strong, highly sensitive, a compact conformation, efficiency advantages of higher.

It is mainly detection, identification and the location that the uhf electromagnetic wave by exciting in detection power equipment shelf depreciation situation is realized local discharge signal that transformer station's local discharge superhigh frequency detects rule, but may there is corona discharge while operation due to the high-tension apparatus in transformer station, on ultra-high frequency signal, stack is disturbed, and the interference of background white noise, cause the starting point of ultra-high frequency signal to be difficult to determine, affected precision and the reliability of high-frequency local discharging location.

Summary of the invention

Based on this, the invention provides a kind of localization method and system of transformer station partial discharge signals, can improve the accuracy and reliability of transformer station partial discharge signals location.

A localization method for transformer station partial discharge signals, comprises the steps:

Gather by least four uhf sensors the local discharge signal that transformer station produces;

Utilize default multiple wavelet function to carry out denoising to described local discharge signal;

Utilize default higher order cumulants flow function to calculate local discharge signal after the described denoising time delay at the described uhf sensor of difference;

According to described time delay and the installation site of uhf sensor described in each, the generation position of calculating described local discharge signal.

A positioning system for transformer station partial discharge signals, comprising:

Acquisition module, for gathering by least four uhf sensors the local discharge signal that transformer station produces;

Denoising module, for utilizing default multiple wavelet function to carry out denoising to described local discharge signal;

Time-delay calculation module, for utilizing default higher order cumulants flow function to calculate local discharge signal after the described denoising time delay at the described uhf sensor of difference;

Position computation module, for according to described time delay and the installation site of uhf sensor described in each, the generation position of calculating described local discharge signal.

The localization method of above-mentioned transformer station partial discharge signals and system, adopt uhf sensor to gather local discharge signal, utilize wavelet transformation to carry out denoising to local discharge signal, the effectively interference of cancelling noise to the identification of signal starting point, improve the precision of follow-up signal delay estimation, and then improved accuracy and the precision of shelf depreciation location.Positional accuracy and the precision of local discharge signal of the present invention are high, and processing speed is fast, have solved the challenge of open type transformer station ultrahigh frequency method local discharge signal location in conventional art, are convenient to practical application.

Brief description of the drawings

Fig. 1 is the schematic flow sheet in one embodiment of localization method of transformer station partial discharge signals of the present invention.

Fig. 2 is the schematic flow sheet of step S12 in Fig. 1.

Fig. 3 is the schematic flow sheet of step S221 in Fig. 2.

Fig. 4 is the structural representation in one embodiment of positioning system of transformer station partial discharge signals of the present invention.

Embodiment

Below in conjunction with embodiment and accompanying drawing, the present invention is described in further detail, but embodiments of the present invention are not limited to this.

As shown in Figure 1, be the schematic flow sheet of the localization method of transformer station partial discharge signals of the present invention, comprising:

S11, the local discharge signal producing by least four uhf sensors collection transformer stations;

The local discharge signal that in transformer station, power equipment produces, ultrahigh frequency (UHF) electromagnetic wave has advantages of strong interference immunity, highly sensitive and velocity of propagation is stable; In the present embodiment, adopt uhf sensor to gather local discharge signal, the local discharge signal collecting is ultra-high frequency signal, and its frequency is 300MHz～3GHz; Concrete, at least needing four groups of ultrahigh frequency omnidirectional sensor arraies, sensor array can be arranged on movable supporting frame; After installation, the coordinate system of substation field, as initial point, is set up in certain position of all right substation field, records the installation site coordinate of each group of sensor array.

S12, the default multiple wavelet function of utilization carry out denoising to described local discharge signal;

To the local discharge signal collecting, first important processing is to carry out denoising; In the present embodiment, adopt wavelet transformation to carry out denoising to local discharge signal, small wave converting method is compared to traditional Fourier transform denoising, its time domain and frequency domain have good locality simultaneously, be more suitable in this class high frequency of local discharge signal, faint and be mixed with clutter jump signal detect.

In a preferred embodiment, as shown in Figure 2, step S12 can comprise:

S221, according to described multiple wavelet function, described local discharge signal is carried out to multiple wavelet decomposition, obtain detail signal and the approximate signal of described local discharge signal;

In the time that local discharge signal is decomposed, multiple wavelet function to choose with decomposing the definite of exponent number be the key of carrying out multiple wavelet decomposition;

In a preferred embodiment, the wavelet basis adopting when carrying out wavelet transformation does not have uniqueness, and different wavelet basiss have different attributes, difference is larger, can select most suitable wavelet basis function according to the feature of local discharge signal, as shown in Figure 3, step S221 can comprise:

S2211, each wavelet basis function to default wavelet function data centralization, calculate the similarity of waveform and the wavelet basis function of described local discharge signal according to following formula:

$r=\frac{\mathrm{\Σ}(s-\stackrel{\‾}{s})(w-\stackrel{\‾}{w})}{\sqrt{\mathrm{\Σ}{(s-\stackrel{\‾}{s})}^2{(w-\stackrel{\‾}{w})}^2}}$

Wherein, r is described similarity, and s is described local discharge signal, and w is the wavelet basis function of described wavelet function data centralization, with be respectively the average of s and w;

In the present embodiment, in the time choosing wavelet function, need to carry out similarity degree comparison to wavelet shapes and signal waveform, choose wavelet function that similarity the is high basis function as wavelet transformation;

Wavelet function data set includes default multiple wavelet basis functions, for example, have the wavelet functions such as Haar wavelet basis, db series wavelet basis, the little wave system of Biorthogonal (biorNr.Nd), the little wave system of Coiflet (coifN), the little wave system of SymletsA (symN), Molet (morl) small echo, Mexican Hat (mexh) small echo, Meyer small echo; Default wavelet basis function is more, optionally gets the wavelet basis function more similar to local discharge signal waveform, thereby improves denoising effect.

S2212, real part using wavelet basis function the highest similarity as multiple wavelet function, carry out the real part of described multiple wavelet function Hilbert transform and obtain the imaginary part of multiple wavelet function, combines described real part and imaginary part and obtain described multiple wavelet function;

Choose after wavelet basis function, need wavelet basis function to form multiple wavelet function, then local discharge signal is decomposed.

S2213, according to described multiple wavelet function, described local discharge signal is carried out to multiple wavelet decomposition, obtain multilayer detail signal and one deck approximate signal of described local discharge signal, wherein, the number of plies of described detail signal is $L_1=\mathrm{fix}\left({\log }_2\right[\frac{l_s}{l_w-1}\left]\right)$ With $L_2=\mathrm{rou}\left({\log }_2\right[\frac{f_s}{f_l}\left]\right)-1$ In smaller value, l _sfor the length of described local discharge signal, l _wfor the described length of the wave filter of wavelet function again, fix represents the value fractions omitted position after calculating to round; f _sfor the sample frequency of described local discharge signal, f _lfor the frequency range minimum value of described local discharge signal, rou rounds after representing the value after calculating to round up.

S222, employing threshold filter carry out denoising to described detail signal;

Local discharge signal carries out resolving into multilayer detail signal and one deck approximate signal after multiple wavelet decomposition, in detail signal, comprise less noise information and a large amount of high-frequency local discharging pulse informations, the detail signal of every one deck is carried out to threshold filter, can reduce the impact of signal noise, therefore threshold value choose particularly crucial; Wavelet threshold filtering is to utilize wavelet transformation, removes wavelet coefficient by a small margin, shrinks or the larger wavelet coefficient of reservation amplitude the noise in Inhibitory signal.When filtering, the definite of threshold value is the key issue of wavelet threshold filtering, directly determining filter effect.

In a preferred embodiment, step S222 can be:

To detail signal described in every one deck, the wavelet coefficient zero setting of threshold value will be less than in described detail signal; Wherein, described threshold value is t _hfor described threshold value, N is the length of wavelet coefficient in every one deck detail signal, and m is the intermediate value of wavelet coefficient in described detail signal.

S223, described detail signal and described approximate signal after utilizing described multiple wavelet function to denoising are reconstructed, and obtain the local discharge signal after described denoising;

In the present embodiment, utilize wavelet transformation to recombinate to the signal after threshold filter, obtain the local discharge signal after denoising; Signal reconstruction is the inverse process of signal wavelet decomposition, local discharge signal carries out being decomposed into multilayer detail signal and one deck approximate signal after wavelet transform process, signal denoising process is processed detail signal, after finishing dealing with, the recycling wavelet function identical with decomposable process is reconstructed, multilayer detail signal after treatment and one deck approximate signal are reformulated to local discharge signal, thereby realize the denoising of local discharge signal.

S13, utilize default higher order cumulants flow function to calculate local discharge signal after the described denoising time delay at the described uhf sensor of difference;

In the present embodiment, to any two local discharge signals that sensor collects, can be expressed as:

$\left\{\begin{array}{c}x\left(t\right)=s\left(t\right)+n_1\left(t\right)\\ y\left(t\right)=\mathrm{As}(t-D)+n_2\left(t\right)\end{array}\right.$

In formula, x (t) and y (t) is respectively sensor 1 and sensor 2 collects shelf depreciation discharge signal at moment t; S (t) is the original local discharge signal of x (t), n ₁(t) noise signal collecting for sensor 1;

Due to the installation site difference of different sensors, signal s (t) has certain delay in the time arriving sensor 2, the original local discharge signal through postponing that As (t-D) receives for sensor 2, n ₂(t) noise signal collecting for sensor 2; Time-delay calculation postpones D estimated time according to observation signal x (t) and y (t);

Solve signal s ₁and s (t) ₂(t) correlativity between, when optimum matching occurs in displacement and is D.

Signal s ₁and s (t) ₂(t) relevance function is:

$\begin{array}{c}{C}_{\mathrm{xy}}\left(\mathrm{\τ}\right)=E\left(x\left(t\right)y(t+\mathrm{\τ})\right)\\ =A{C}_{2s}(\mathrm{\τ}-D)\end{array}$

In formula, C _xy(τ) be the relevance function of signal x (t) and signal y (t), τ is time shift variable, and E represents to average, C _2s(τ) be the autocorrelation function of signal s (t); A is default factor;

Noise signal n ₁and n (t) ₂(t) be zero-mean, independent of one another, and independent with original local discharge signal:

C _2s(τ)＝E{s(t)s(t+τ)}

Be C _xy(τ) in the time of τ=D, get peak value.

The Higher Order Cumulants of two observation signals is carried out to maximum matching value tracking, realize the time delay of signal and estimate to calculate, the present embodiment is chosen 4 rank semi-invariant C _xxyy(τ ₁, τ ₂, τ ₃) carry out time delay estimation;

Wherein, C _xxyy(τ ₁, τ ₂, τ ₃)=C ₄{ x (t), x ^*(t+ τ ₁), y (t+ τ ₂), y ^*(t+ τ ₃), * represents conjugation, τ ₁, τ ₂and τ ₃represent time delay.

S14, according to described time delay and the installation site of uhf sensor described in each, the generation position of calculating described local discharge signal;

In the present embodiment, four signal detection sensors of minimum needs are located the generation position of local discharge signal, and time and the position of setting local discharge signal generation are respectively t ₀(x ₀, y ₀, z ₀); The installation site of each sensor array is (x _i, y _i, z _i), i=1,2,3,4; Sensor detects that the initial time of local signal is t _i0, i=1,2,3,4.:

$\left\{\begin{array}{c}c\×(t_{10}-t_0)=\sqrt{{(x_1-x_0)}^2+{(y_1-y_0)}^2+{(z_1-z_0)}^2}\\ c\×(t_{20}-t_0)=\sqrt{{(x_2-x_0)}^2+{(y_2-y_0)}^2+{(z_2-z_0)}^2}\\ c\×(t_{30}-t_0)=\sqrt{{(x_3-x_0)}^2+{(y_3-y_0)}^2+{(z_3-z_0)}^2}\\ c\×(t_{40}-t_0)=\sqrt{{(x_4-x_0)}^2+{(y_4-y_0)}^2+{(z_4-z_0)}^2}\end{array}\right.$

Wherein, c is shelf depreciation high-frequency signal velocity of propagation.

The time delay t that between known two sensors, local discharge signal reaches ₁, t ₂, t ₃, and the position (x of known detection antenna placement _i, y _i, z _i), i=1,2,3,4.Pass through formula

$\left\{\begin{array}{c}c\×t_1=c\×(t_{20}-t_{10})=\sqrt{{(x_1-x_0)}^2+{(y_1-y_0)}^2+{(z_1-z_0)}^2}-\sqrt{{(x_2-x_0)}^2+{(y_2-y_0)}^2+{(z_2-z_0)}^2}\\ c\×t_2=c\×(t_{30}-t_{20})=\sqrt{{(x_2-x_0)}^2+{(y_2-y_0)}^2+{(z_2-z_0)}^2}-\sqrt{{(x_3-x_0)}^2+{(y_3-y_0)}^2+{(z_3-z_0)}^2}\\ c\×t_3=c\×(t_{40}-t_{30})=\sqrt{{(x_3-x_0)}^2+{(y_3-y_0)}^2+{(z_3-z_0)}^2}-\sqrt{{(x_4-x_0)}^2+{(y_4-y_0)}^2+{(z_4-z_0)}^2}\end{array}\right.$

Can solve the position (x that local discharge signal produces ₀, y ₀, z ₀).

As shown in Figure 4, the present invention also provides a kind of positioning system of transformer station partial discharge signals, comprising:

Acquisition module 41, for gathering by least four uhf sensors the local discharge signal that transformer station produces;

The local discharge signal that in transformer station, power equipment produces, ultrahigh frequency (UHF) electromagnetic wave has advantages of strong interference immunity, highly sensitive and velocity of propagation is stable; In the present embodiment, adopt uhf sensor to gather local discharge signal, the Partial discharge signal collecting is ultra-high frequency signal, and its frequency is 300MHz～3GHz; Concrete, at least needing four groups of ultrahigh frequency omnidirectional sensor arraies, sensor array can be arranged on movable supporting frame; After installation, the coordinate system of substation field, as initial point, is set up in certain position of all right substation field, records the installation site coordinate of each group of sensor array.

Denoising module 42, for utilizing default multiple wavelet function to carry out denoising to described local discharge signal;

To the local discharge signal collecting, first important processing is to carry out denoising; In the present embodiment, adopt wavelet transformation to carry out denoising to local discharge signal, small wave converting method is compared to traditional Fourier transform denoising, its time domain and frequency domain have good locality simultaneously, be more suitable in this class high frequency of local discharge signal, faint and be mixed with clutter jump signal detect.

In a preferred embodiment, denoising module 41 can comprise:

Multiple wavelet decomposition module, for according to described multiple wavelet function, described local discharge signal being carried out to multiple wavelet decomposition, obtains detail signal and the approximate signal of described local discharge signal;

In the time that local discharge signal is decomposed, multiple wavelet function to choose with decomposing the definite of exponent number be the key of carrying out multiple wavelet decomposition;

In a preferred embodiment, the wavelet basis adopting when carrying out wavelet transformation does not have uniqueness, and different wavelet basiss have different attributes, and difference is larger, can select most suitable wavelet basis function according to the feature of local discharge signal, multiple wavelet decomposition module can comprise:

Similarity calculation module, for each wavelet basis function of the wavelet function data centralization to default, calculate the similarity of waveform and the wavelet basis function of described local discharge signal according to following formula:

$r=\frac{\mathrm{\Σ}(s-\stackrel{\‾}{s})(w-\stackrel{\‾}{w})}{\sqrt{\mathrm{\Σ}{(s-\stackrel{\‾}{s})}^2{(w-\stackrel{\‾}{w})}^2}}$

Wherein, r is described similarity, and s is described local discharge signal, and w is the wavelet basis function of described wavelet function data centralization, with be respectively the average of s and w;

In the present embodiment, in the time choosing wavelet function, conventionally wavelet shapes and signal waveform are carried out to similarity degree comparison, choose wavelet function that similarity the is high basis function as wavelet transformation;

Wavelet function data set includes default multiple wavelet basis functions, for example, have the wavelet functions such as Haar wavelet basis, db series wavelet basis, the little wave system of Biorthogonal (biorNr.Nd), the little wave system of Coiflet (coifN), the little wave system of SymletsA (symN), Molet (morl) small echo, Mexican Hat (mexh) small echo, Meyer small echo; Default wavelet basis function is more, optionally gets the wavelet basis function more similar to local discharge signal waveform, thereby improves denoising effect.

Composite module, for the real part using wavelet basis function the highest similarity as multiple wavelet function, carries out the real part of described multiple wavelet function Hilbert transform and obtains the imaginary part of multiple wavelet function, combines described real part and imaginary part and obtains described multiple wavelet function;

Choose after wavelet basis function, need wavelet basis function to form multiple wavelet function, then local discharge signal is decomposed.

Signal decomposition module, for according to described multiple wavelet function, described local discharge signal being carried out to multiple wavelet decomposition, obtains multilayer detail signal and one deck approximate signal of described local discharge signal, and wherein, the number of plies of described detail signal is $L_1=\mathrm{fix}\left({\log }_2\right[\frac{l_s}{l_w-1}\left]\right)$ With $L_2=\mathrm{rou}\left({\log }_2\right[\frac{f_s}{f_l}\left]\right)-1$ In smaller value, l _sfor the length of described local discharge signal, l _wfor the described length of the wave filter of wavelet function again, fix represents the value fractions omitted position after calculating to round; f _sfor the sample frequency of described local discharge signal, f _lfor the frequency range minimum value of described local discharge signal, rou rounds after representing the value after calculating to round up.

Filtration module, for adopting threshold filter to carry out denoising to described detail signal;

Local discharge signal carries out resolving into multilayer detail signal and one deck approximate signal after multiple wavelet decomposition, in detail signal, comprise less noise information and a large amount of high-frequency local discharging pulse informations, the detail signal of every one deck is carried out to threshold filter, can reduce the impact of signal noise, therefore threshold value choose particularly crucial; Wavelet threshold filtering is to utilize wavelet transformation, removes wavelet coefficient by a small margin, shrinks or the larger wavelet coefficient of reservation amplitude the noise in Inhibitory signal.When filtering, the definite of threshold value is the key issue of wavelet threshold filtering, directly determining filter effect.

In a preferred embodiment, described filtration module also for: to detail signal described in every one deck, will in described detail signal, be less than the wavelet coefficient zero setting of threshold value; Wherein, described threshold value is t _hfor described threshold value, N is the length of wavelet coefficient in every one deck detail signal, and m is the intermediate value of wavelet coefficient in described detail signal.

Reconstructed module, is reconstructed for described detail signal and described approximate signal after utilizing described multiple wavelet function to denoising, obtains the local discharge signal after described denoising;

In the present embodiment, utilize wavelet transformation to recombinate to the signal after threshold filter, obtain the local discharge signal after denoising; Signal reconstruction is the inverse process of signal wavelet decomposition, local discharge signal carries out being decomposed into multilayer detail signal and one deck approximate signal after wavelet transform process, signal denoising process is processed detail signal, after finishing dealing with, the recycling wavelet function identical with decomposable process is reconstructed, multilayer detail signal after treatment and one deck approximate signal are reformulated to local discharge signal, thereby realize the denoising of local discharge signal.

Time-delay calculation module 43, for utilizing default higher order cumulants flow function to calculate local discharge signal after the described denoising time delay at the described uhf sensor of difference;

In the present embodiment, to any two local discharge signals that sensor collects, can be expressed as:

$\left\{\begin{array}{c}x\left(t\right)=s\left(t\right)+n_1\left(t\right)\\ y\left(t\right)=\mathrm{As}(t-D)+n_2\left(t\right)\end{array}\right.$

In formula, x (t) and y (t) is respectively sensor 1 and sensor 2 collects shelf depreciation discharge signal at moment t; S (t) is the original local discharge signal of x (t), n ₁(t) noise signal collecting for sensor 1;

Due to the installation site difference of different sensors, signal s (t) has certain delay in the time arriving sensor 2, the original local discharge signal through postponing that As (t-D) receives for sensor 2, n ₂(t) noise signal collecting for sensor 2; Time delay is estimated to postpone D estimated time according to observation signal x (t) and y (t);

Solve signal s ₁and s (t) ₂(t) correlativity between, when optimum matching occurs in displacement and is D.

Signal s ₁and s (t) ₂(t) relevance function is:

$\begin{array}{c}{C}_{\mathrm{xy}}\left(\mathrm{\τ}\right)=E\left(x\left(t\right)y(t+\mathrm{\τ})\right)\\ =A{C}_{2s}(\mathrm{\τ}-D)\end{array}$

In formula, C _xy(τ) be the cross correlation function of signal x (t) and signal y (t), τ is time shift variable, and E represents to average, C _2s(τ) be the autocorrelation function of signal s (t); A is default factor;

Noise signal n ₁and n (t) ₂(t) be zero-mean, independent of one another, and independent with original local discharge signal,

C _2s(τ)＝E{s(t)s(t+τ)}

Be C _xy(τ) in the time of τ=D, get peak value.

The Higher Order Cumulants of two observation signals is carried out to maximum matching value tracking, realize the time delay of signal and estimate to calculate, the present embodiment is chosen 4 rank semi-invariant C _xxyy(τ ₁, τ ₂, τ ₃) carry out time delay estimation;

Wherein, C _xxyy(τ ₁, τ ₂, τ ₃)=C ₄{ x (t), x ^*(t+ τ ₁), y (t+ τ ₂), y ^*(t+ τ ₃), * represents conjugation, τ ₁, τ ₂and τ ₃represent time delay.

Position computation module 44, for according to described time delay and the installation site of uhf sensor described in each, the generation position of calculating described local discharge signal;

In the present embodiment, four signal detection sensors of minimum needs are located the generation position of local discharge signal, and time and the position of setting local discharge signal generation are respectively t ₀(x ₀, y ₀, z ₀); The installation site of each sensor array is (x _i, y _i, z _i), i=1,2,3,4; Sensor detects that the initial time of local signal is t _i0, i=1,2,3,4.:

$\left\{\begin{array}{c}c\×(t_{10}-t_0)=\sqrt{{(x_1-x_0)}^2+{(y_1-y_0)}^2+{(z_1-z_0)}^2}\\ c\×(t_{20}-t_0)=\sqrt{{(x_2-x_0)}^2+{(y_2-y_0)}^2+{(z_2-z_0)}^2}\\ c\×(t_{30}-t_0)=\sqrt{{(x_3-x_0)}^2+{(y_3-y_0)}^2+{(z_3-z_0)}^2}\\ c\×(t_{40}-t_0)=\sqrt{{(x_4-x_0)}^2+{(y_4-y_0)}^2+{(z_4-z_0)}^2}\end{array}\right.$

Wherein, c is shelf depreciation high-frequency signal velocity of propagation.

The time delay t that between known two sensors, local discharge signal reaches ₁, t ₂, t ₃, and the position (x of known detection antenna placement _i, y _i, z _i), i=1,2,3,4.Pass through formula

$\left\{\begin{array}{c}c\×t_1=c\×(t_{20}-t_{10})=\sqrt{{(x_1-x_0)}^2+{(y_1-y_0)}^2+{(z_1-z_0)}^2}-\sqrt{{(x_2-x_0)}^2+{(y_2-y_0)}^2+{(z_2-z_0)}^2}\\ c\×t_2=c\×(t_{30}-t_{20})=\sqrt{{(x_2-x_0)}^2+{(y_2-y_0)}^2+{(z_2-z_0)}^2}-\sqrt{{(x_3-x_0)}^2+{(y_3-y_0)}^2+{(z_3-z_0)}^2}\\ c\×t_3=c\×(t_{40}-t_{30})=\sqrt{{(x_3-x_0)}^2+{(y_3-y_0)}^2+{(z_3-z_0)}^2}-\sqrt{{(x_4-x_0)}^2+{(y_4-y_0)}^2+{(z_4-z_0)}^2}\end{array}\right.$

Can solve the position (x that local discharge signal produces ₀, y ₀, z ₀).

The localization method of transformer station partial discharge signals of the present invention and system, adopt uhf sensor to gather local discharge signal, utilize wavelet transformation to carry out denoising to local discharge signal, the effectively interference of cancelling noise to the identification of signal starting point, improve the precision of follow-up signal delay estimation, and then improved accuracy and the precision of shelf depreciation location.Positional accuracy and the precision of local discharge signal of the present invention are high, and processing speed is fast, have solved the challenge of open type transformer station ultrahigh frequency method local discharge signal location in conventional art, are convenient to practical application.

The above embodiment has only expressed several embodiment of the present invention, and it describes comparatively concrete and detailed, but can not therefore be interpreted as the restriction to the scope of the claims of the present invention.It should be pointed out that for the person of ordinary skill of the art, without departing from the inventive concept of the premise, can also make some distortion and improvement, these all belong to protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.

Claims (8)

1. a localization method for transformer station partial discharge signals, is characterized in that, comprises the steps:

Gather by least four uhf sensors the local discharge signal that transformer station produces;

Utilize default multiple wavelet function to carry out denoising to described local discharge signal;

Utilize default higher order cumulants flow function to calculate local discharge signal after the described denoising time delay at the described uhf sensor of difference;

According to described time delay and the installation site of uhf sensor described in each, the generation position of calculating described local discharge signal.

2. the localization method of transformer station partial discharge signals according to claim 1, is characterized in that, the described step of utilizing described multiple wavelet function to carry out denoising to described local discharge signal comprises:

According to described multiple wavelet function, described local discharge signal is carried out to multiple wavelet decomposition, obtain detail signal and the approximate signal of described local discharge signal;

Adopt threshold filter to carry out denoising to described detail signal;

Described detail signal and described approximate signal after utilizing described multiple wavelet function to denoising are reconstructed, and obtain the local discharge signal after described denoising.

3. the localization method of transformer station partial discharge signals according to claim 2, it is characterized in that, describedly according to described multiple wavelet function, described local discharge signal is carried out to multiple wavelet decomposition, obtains the detail signal of described local discharge signal and the step of approximate signal comprises:

To each wavelet basis function of default wavelet function data centralization, calculate the similarity of waveform and the wavelet basis function of described local discharge signal according to following formula:

$r=\frac{\mathrm{\Σ}(s-\stackrel{\‾}{s})(w-\stackrel{\‾}{w})}{\sqrt{\mathrm{\Σ}{(s-\stackrel{\‾}{s})}^2{(w-\stackrel{\‾}{w})}^2}}$

Wherein, r is described similarity, and s is described local discharge signal, and w is the wavelet basis function of described wavelet function data centralization, with be respectively the average of s and w;

Real part using wavelet basis function the highest similarity as multiple wavelet function, carries out the real part of described multiple wavelet function Hilbert transform and obtains the imaginary part of multiple wavelet function, combines described real part and imaginary part and obtains described multiple wavelet function;

According to described multiple wavelet function, described local discharge signal is carried out to multiple wavelet decomposition, obtain multilayer detail signal and one deck approximate signal of described local discharge signal, wherein, the number of plies of described detail signal is $L_1=\mathrm{fix}\left({\log }_2\right[\frac{l_s}{l_w-1}\left]\right)$ With $L_2=\mathrm{rou}\left({\log }_2\right[\frac{f_s}{f_l}\left]\right)-1$ In smaller value, l _sfor the length of described local discharge signal, l _wfor the described length of the wave filter of wavelet function again, fix represents the value fractions omitted position after calculating to round; f _sfor the sample frequency of described local discharge signal, f _lfor the frequency range minimum value of described local discharge signal, rou rounds after representing the value after calculating to round up.

4. the localization method of transformer station partial discharge signals according to claim 3, is characterized in that, the step that described employing threshold filter carries out denoising to described detail signal comprises:

To detail signal described in every one deck, the wavelet coefficient zero setting of threshold value will be less than in described detail signal; Wherein, described threshold value is n is the length of wavelet coefficient in every one deck detail signal, and m is the intermediate value of wavelet coefficient in described detail signal.

5. a positioning system for transformer station partial discharge signals, is characterized in that, comprising:

Acquisition module, for gathering by least four uhf sensors the local discharge signal that transformer station produces;

Denoising module, for utilizing default multiple wavelet function to carry out denoising to described local discharge signal;

Time-delay calculation module, for utilizing default higher order cumulants flow function to calculate local discharge signal after the described denoising time delay at the described uhf sensor of difference;

Position computation module, for according to described time delay and the installation site of uhf sensor described in each, the generation position of calculating described local discharge signal.

6. the positioning system of transformer station partial discharge signals according to claim 5, is characterized in that, described denoising module comprises:

Multiple wavelet decomposition module, for according to described multiple wavelet function, described local discharge signal being carried out to multiple wavelet decomposition, obtains detail signal and the approximate signal of described local discharge signal;

Filtration module, for adopting threshold filter to carry out denoising to described detail signal;

Reconstructed module, is reconstructed for described detail signal and described approximate signal after utilizing described multiple wavelet function to denoising, obtains the local discharge signal of denoising.

7. the positioning system of transformer station partial discharge signals according to claim 6, is characterized in that, described decomposing module comprises:

Similarity calculation module, for each wavelet basis function of the wavelet function data centralization to default, calculate the similarity of waveform and the wavelet basis function of described local discharge signal according to following formula:

$r=\frac{\mathrm{\Σ}(s-\stackrel{\‾}{s})(w-\stackrel{\‾}{w})}{\sqrt{\mathrm{\Σ}{(s-\stackrel{\‾}{s})}^2{(w-\stackrel{\‾}{w})}^2}}$

Wherein, r is described similarity, and s is described local discharge signal, and w is the wavelet basis function of described wavelet function data centralization, with be respectively the average of s and w;

Composite module, for the real part using wavelet basis function the highest similarity as multiple wavelet function, carries out the real part of described multiple wavelet function Hilbert transform and obtains the imaginary part of multiple wavelet function, combines described real part and imaginary part and obtains described multiple wavelet function;

Signal decomposition module, for according to described multiple wavelet function, described local discharge signal being carried out to multiple wavelet decomposition, obtains multilayer detail signal and one deck approximate signal of described local discharge signal, and wherein, the number of plies of described detail signal is $L_1=\mathrm{fix}\left({\log }_2\right[\frac{l_s}{l_w-1}\left]\right)$ With $L_2=\mathrm{rou}\left({\log }_2\right[\frac{f_s}{f_l}\left]\right)-1$ In smaller value, l _sfor the length of described local discharge signal, l _wfor the described length of the wave filter of wavelet function again, fix represents the value fractions omitted position after calculating to round; f _sfor the sample frequency of described local discharge signal, f _lfor the frequency range minimum value of described local discharge signal, rou rounds after representing the value after calculating to round up.

8. the positioning system of transformer station partial discharge signals according to claim 7, is characterized in that, described filtration module also for: to detail signal described in every one deck, will in described detail signal, be less than the wavelet coefficient zero setting of described threshold value; Wherein, described threshold value is n is the length of wavelet coefficient in every one deck detail signal, and m is the intermediate value of described detail signal wavelet coefficient.