CN112269095A - Fault detection method based on fault current intermittent reignition and extinguishment characteristics - Google Patents

Abstract

The invention discloses a fault detection method based on fault current intermittent reignition and extinguishment characteristics, which is implemented according to the following steps: carrying out mode-changing decomposition on the zero-sequence current to obtain each intrinsic mode, calculating kurtosis values of each intrinsic mode, and defining the mode corresponding to the maximum kurtosis value as a characteristic mode; calculating a Teager-Kaiser energy operator for the characteristic mode, and carrying out interval division on the waveform of the energy operator; calculating the time entropy value of each interval to finally obtain a total time entropy value; and when the total time entropy is not 0, judging that the power distribution network has a high-resistance fault, and when the total time entropy is equal to 0, judging that the power distribution network has capacitor switching or load switching. The invention aims to distinguish 3 working conditions of high-resistance fault, capacitance switching and load switching in a power distribution network, and accurately detect the working condition of the high-resistance fault to ensure that a relay protection device operates to trip; and for the capacitor switching and load switching working conditions, a signal is not sent to the relay protection device.

Description

Fault detection method based on fault current intermittent reignition and extinguishment characteristics

Technical Field

The invention belongs to the technical field of relay protection of a power distribution network of a power system, and particularly relates to a fault detection method based on fault current intermittent reignition and extinguishment characteristics.

Background

Under the influence of power transmission corridors, natural environment, low overhead line distance and other factors, single-phase high-resistance grounding faults of non-ideal conductors, such as wire falling on concrete ground, grassland, sand and the like, often occur in the power distribution network. According to incomplete statistics, high resistance ground faults account for about 5% -10% of the total number of ground faults. High resistance earth fault current is weak, and the fault current is less than 10% of load current even according to different fault point media, so that the traditional relay protection method cannot accurately detect High resistance fault (HIF), further fault deterioration is caused, then serious fault is developed, and safe operation of a power distribution network is damaged. Aiming at the problem of detecting the high-resistance fault of the power distribution network, a large amount of research is carried out by scholars at home and abroad, and a plurality of detection methods of different types are provided, but in conclusion, the existing high-resistance fault detection method still has the following problems:

1) and (3) fault feature extraction: most of the basis functions of the existing methods such as Fourier transform, wavelet transform, S transform and the like are fixed, so that the characteristic characterization capability is not strong, the method has no self-adaptability in the extraction process, the characteristic component without actual physical significance is easy to obtain, and the method is not beneficial to the construction of identification criteria. Although the empirical mode division algorithm has an adaptive characteristic, the decomposition of the empirical mode division algorithm is easy to generate the problems of mode aliasing and end point effect.

2) The detection criterion construction aspect is as follows: most of the existing methods construct network discrimination methods such as a neural network, a support vector machine, a decision tree and the like on the basis of obtaining characteristic components, and although the methods have high judgment accuracy, the methods need a sample library in advance and need to be trained, so that the judgment time is long in practical application.

3) The essential difference of high-resistance fault, capacitor switching and load switching on fault current waveform is ignored: when HIF occurs, intermittent reignition and extinguishment phenomena often occur at a zero crossing point of a fault arc, and when Capacitor Switching (CS) and Load Switching (LS) occur, the power distribution network is impacted instantly, but the intermittent reignition and extinguishment phenomena cannot occur, so that an identification criterion can be constructed by using the characteristic, and the HIF, CS and LS can be accurately identified.

Disclosure of Invention

The invention aims to provide a fault detection method based on fault current intermittent reignition and extinguishment characteristics, which can accurately judge the fault type.

The invention adopts the technical scheme that a fault detection method based on the intermittent reignition and extinguishment characteristics of fault current is implemented according to the following steps:

step 1, carrying out mode-changing decomposition on zero-sequence current to obtain each intrinsic mode, calculating kurtosis values of each intrinsic mode, and defining the mode corresponding to the maximum kurtosis value as a characteristic mode;

step 2, calculating Teager-Kaiser energy operators for the characteristic modes, and carrying out interval division on the waveforms of the energy operators;

step 3, calculating the time entropy value of each interval to finally obtain the total time entropy value;

and 4, when the total time entropy is not 0, judging that the power distribution network has a high-resistance fault, and when the total time entropy is equal to 0, judging that the power distribution network has capacitor switching or load switching.

The specific process of the step 1 is as follows: respectively decomposing the zero sequence current of the power distribution network by adopting a variable modal decomposition algorithm to obtain intrinsic modal components IMF₁(n)，IMF₂(n)，…，IMF_s(n), s is a mode number, and n is a sampling point; for IMF₁(n)，IMF₂(n)，…，IMF_s(n) separately calculating kurtosis values k₁，k₂，…，k_s(ii) a At kurtosis value k₁，k₂，…，k_sTo find out the maximum kurtosis value k_tThen the maximum kurtosis value k_tCorresponding intrinsic mode component IMF_tAnd (n) is defined as a characteristic mode of the zero sequence current of the power distribution network.

Modal component IMF_sKurtosis value k of (n)_sThe calculation formula is as follows:

wherein, mu and sigma are intrinsic mode components IMF_sAnd (N) the mean value and the standard deviation, wherein N is the total number of sampling points.

The specific process of the step 2 is as follows: for characteristic mode IMF_t(n) calculating Teager-Kaiser energy operator TKEO_t(n); to energy operator TKEO_tThe waveform of (n) is divided into intervals at intervals, and the interval Δ T of the interval division is 0.01s, thereby obtaining an interval 1: TKEO_{t_1}(n), interval 2: TKEO_{t_2}(n), …, interval x: TKEO_{t_x}(n), x is the interval number.

For characteristic mode IMF_t(n) calculating Teager-Kaiser energy operator TKEO_tThe calculation formula of (n) is as follows:

TKEO_t(n)＝[IMF_t(n)]²-IMF_t(n+1)IMF_t(n-1)

in the above formula, IMF_t(n)，IMF_t(n+1)，IMF_tAnd (n-1) is the characteristic mode corresponding to the nth, the (n + 1) th and the (n-1) th sampling points respectively.

The specific process of the step 3 is as follows: for each interval TKEO_{t_1}(n)，TKEO_{t_2}(n)，…，TKEO_{t_x}(n) separately calculating the time entropy values TS_{t_1}，TS_{t_2}，…，TS_{t_x}(ii) a For time entropy values TS of intervals_{t_1}，TS_{t_2}，…，TS_{t_x}And adding and summing to obtain a total time entropy value TS.

Interval TKEO_{t_x}(n) the calculation of the time entropy value is as follows:

TS_{t_x}＝-∑p_{t_x}(m)[lnp_{t_x}(m)]

where m is the sampling point in the interval x, p_{t_x}(m) is the interval x energy operator TKEO_{t_x}(n) the probability that the number of peaks accounts for the number of peaks in the total interval;

the total time entropy value TS is calculated as follows:

wherein l is the total number of intervals.

The invention has the beneficial effects that:

1) the method adopts a variable modal decomposition algorithm with self-adaptive characteristics to extract the characteristics of the transient zero-sequence current, and a series of intrinsic modal components are obtained; a kurtosis calculation formula is introduced, a mode corresponding to the maximum kurtosis value is selected to be defined as a characteristic mode, the accurate extraction of transient zero sequence current characteristics is realized, a high-frequency component reflecting sensitive change is obtained, and a foundation is laid for the construction of the next detection criterion.

2) The detection criterion construction aspect is as follows: aiming at the defects of large calculated amount and the like of the existing network identification method, a Teager-Kaiser energy operator calculation characteristic mode is provided, and the Teager-Kaiser energy operator during high-resistance fault, capacitor switching and load switching is obtained; further, a time entropy calculation theory is constructed by carrying out interval division on the Teager-Kaiser energy operator, and total time entropy values of HIF, CS and LS are obtained; finally, by identifying whether the total time entropy is 0, accurate detection of HIF is achieved.

3) The intermittent reignition and extinguishing characteristics of the high-resistance fault current are fully utilized: the intermittent re-ignition and extinguishing phenomena of HIF current waveforms near zero crossing points are fully utilized, the CS and LS switching are instantaneous disturbance, zero sequence current waveforms of the HIF current waveforms show oscillation attenuation trend, a characteristic mode reflecting the law is accurately extracted by adopting a VMD method, and the HIF is accurately identified through Teager-Kaiser energy operator and time entropy calculation.

Drawings

FIG. 1 is a flow chart of a fault detection method based on fault current intermittent reignition and extinguishment features in accordance with the present invention;

FIG. 2 is a schematic diagram of a 10kV radial distribution network according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of zero sequence current when HIF occurs in an embodiment of the invention;

FIG. 4 is a schematic diagram of zero sequence current when CS occurs according to an embodiment of the present invention;

fig. 5 is a schematic diagram of zero sequence current when LS occurs in the embodiment of the present invention.

Detailed Description

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

The invention relates to a fault detection method based on fault current intermittent reignition and extinguishment characteristics, which is specifically implemented according to the following steps as shown in figure 1:

step 1, carrying out mode-changing decomposition on zero-sequence current to obtain each intrinsic mode, calculating kurtosis values of each intrinsic mode, and defining the mode corresponding to the maximum kurtosis value as a characteristic mode;

the specific process is as follows: respectively decomposing the zero sequence current of the power distribution network by adopting a variable modal decomposition algorithm to obtain intrinsic modal components IMF₁(n)，IMF₂(n)，…，IMF_s(n), s is a mode number, and n is a sampling point; for IMF₁(n)，IMF₂(n)，…，IMF_s(n) separately calculating kurtosis values k₁，k₂，…，k_s(ii) a At kurtosis value k₁，k₂，…，k_sTo find out the maximum kurtosis value k_tThen the maximum kurtosis value k_tCorresponding intrinsic mode component IMF_tAnd (n) is defined as a characteristic mode of the zero sequence current of the power distribution network.

Modal component IMF_sKurtosis value k of (n)_sThe calculation formula is as follows:

wherein, mu and sigma are intrinsic mode components IMF_s(N) mean, standard deviation, N isThe number of total sampling points.

Step 2, calculating Teager-Kaiser energy operators for the characteristic modes, and carrying out interval division on the waveforms of the energy operators; the specific process of the step 2 is as follows: for characteristic mode IMF_t(n) calculating Teager-Kaiser energy operator TKEO_t(n); to energy operator TKEO_tThe waveform of (n) is divided into intervals at intervals, and the interval Δ T of the interval division is 0.01s, thereby obtaining an interval 1: TKEO_{t_1}(n), interval 2: TKEO_{t_2}(n), …, interval x: TKEO_{t_x}(n), x is the interval number.

For characteristic mode IMF_t(n) calculating Teager-Kaiser energy operator TKEO_tThe calculation formula of (n) is as follows:

TKEO_t(n)＝[IMF_t(n)]²-IMF_t(n+1)IMF_t(n-1)

in the above formula, IMF_t(n)，IMF_t(n+1)，IMF_tAnd (n-1) is the characteristic mode corresponding to the nth, the (n + 1) th and the (n-1) th sampling points respectively.

Step 3, calculating the time entropy value of each interval to finally obtain the total time entropy value; the specific process of the step 3 is as follows: for each interval TKEO_{t_1}(n)，TKEO_{t_2}(n)，…，TKEO_{t_x}(n) separately calculating the time entropy values TS_{t_1}，TS_{t_2}，…，TS_{t_x}(ii) a For time entropy values TS of intervals_{t_1}，TS_{t_2}，…，TS_{t_x}And adding and summing to obtain a total time entropy value TS.

Interval TKEO_{t_x}(n) the calculation of the time entropy value is as follows:

TS_{t_x}＝-∑p_{t_x}(m)[lnp_{t_x}(m)]

where m is the sampling point in the interval x, p_{t_x}(m) is the interval x energy operator TKEO_{t_x}(n) the probability that the number of peaks accounts for the number of peaks in the total interval;

the total time entropy value TS is calculated as follows:

wherein l is the total number of intervals.

And 4, when the total time entropy is not 0, judging that the power distribution network has a high-resistance fault, and when the total time entropy is equal to 0, judging that the power distribution network has capacitor switching or load switching.

The fault detection method based on the fault current intermittent reignition and extinguishment characteristics has the working principle that:

1. high resistive fault current intermittent reignition extinction feature

When a single-phase grounding high-resistance fault occurs in the power distribution network, the arc current intermittently reignites and extinguishes at the zero crossing point, so that the high-resistance fault is different from capacitor switching and load switching. Compared with the high-resistance fault, when the capacitor is switched and the load is switched, the current appears as instantaneous sudden change, the current waveform tends to be stable along with the ending of the processes of the capacitor switching and the load switching, and further analysis shows that the current waveform during the capacitor switching and the load switching is an oscillation attenuation waveform which is essentially different from the intermittent reignition and extinguishment waveform during the high-resistance fault. Therefore, the characteristic can be fully utilized to construct a high-resistance fault detection criterion, and further high-resistance faults, capacitance switching and load switching can be identified.

2. Decomposition by varying modes

In the theory of Variable Mode Decomposition (VMD), the Intrinsic Mode Function (IMF) is defined as an am-fm signal, which is expressed as:

u_k(t)＝A_k(t)cos[φ_k(t)] (1)

in formula (1): a. the_k(t) is u_k(t) instantaneous amplitude. Omega_k(t) is u_kInstantaneous frequency of (t), ω_k(t)＝φ′_k(t)＝dφ_k(t)/dt。A_k(t) and ω_k(t) relative to the phase phi_k(t) is slowly varying, i.e. at [ t-delta, t + delta]Within a range of intervals of (u)_k(t) can be regarded as an amplitude of A_k(t) frequency ω_k(t) harmonic signals. (wherein δ is 2 π/φ'_k(t))

1) Construction of variation problem

Assuming each "mode" is a finite bandwidth with a center frequency, the variational problem is described as seeking k mode functions u_k(t) minimizing the sum of the estimated bandwidths of each mode, wherein the constraint condition is that the sum of the modes is equal to the input signal f, and the specific construction steps are as follows:

step a: obtaining each mode u through Hilbert conversion_k(t) the analytic signal, in order to obtain its single-sided spectrum:

step b: mixing the analysis signals of each mode to obtain an estimated center frequency

Modulating the spectrum of each mode to a respective fundamental band:

step c: calculating the square L of the gradient of the above demodulated signal²Norm, estimating the bandwidth of each modal signal, and the constrained variation problem is as follows:

wherein, { u_k}＝{u₁,…，u_K}，{ω_k}＝{ω₁,…，ω_K}，

2) Solution of variational problem

Step A: introducing a secondary penalty factor alpha and a Lagrange multiplier lambda (t), and changing the constraint variation problem into an unconstrained variation problem, wherein the secondary penalty factor can ensure the reconstruction precision of a signal under the condition of existence of Gaussian noise, the Lagrange multiplier keeps the constraint condition strict, and the expanded Lagrange expression is as follows:

and B: the VMD adopts a multiplicative operator alternating direction method (ADMM) to solve the variation problem, and alternately updates

And λⁿ⁺¹The "saddle point" of the extended lagrangian expression is sought.

Wherein

The value problem of (a) can be expressed as:

in the formula: omega_kIs equivalent to

Is equivalent to

Transforming equation (6) to the frequency domain using a Parseval/Plancherel Fourier equidistant transform:

using ω of item 1 as ω - ω_kInstead of this, the user can,

converting equation (8) into the form of integration of non-negative frequency bins:

at this time, the solution of the secondary optimization problem is:

according to the same process, the problem of the center frequency is firstly converted into the frequency domain:

the updating method of the solved center frequency comprises the following steps:

in the formula:

corresponding to the current residual amount

Wiener filtering of (1);

is the center of gravity of the current mode function power spectrum; to pair

Performing inverse Fourier transform to obtain real part of u_k(t)}。

The basic steps of the VMD algorithm are as follows:

(1) initialization

And n;

(2)updating u according to equations (11) and (12)_kAnd ω_k；

(3) Updating lambda:

(4) for a given discrimination accuracy e > 0, if

Stopping the iteration, otherwise, returning to the step B.

From the final algorithm, the VMD is very simple, firstly, each mode is directly and continuously updated in a frequency domain, and finally, the VMD is transformed into a time domain through Fourier inversion; second, as the center of gravity of the power spectrum of each mode, the center frequency is estimated again and updated in this cycle.

Teager-Kaiser energy operator

The Teager-Kaiser energy operator is a nonlinear local difference operator, and the specific principle is as follows:

for amplitude A and frequency f, sampling frequency f is adopted_sThe expression for the obtained signal x (n) is:

x(n)＝Acos(Ωn+φ) (13)

wherein, omega is 2 pi f/f_sWhere φ is the initial phase angle, the adjacent sample points of equation (13) may form the following equation:

solving the above equation yields:

A²sin²(Ω)＝x²(n)-x(n-1)x(n+1) (15)

when the omega is sufficiently small, the flow rate of the gas,

when Ω < π/4, the numerical error of sin Ω and Ω is less than 11%, and therefore, equation (15) can be written as:

A²Ω²＝x²(n)-x(n+1)x(n-1) (16)

the Teager-Kaiser energy operator defining signal x (n) is:

ψ[x(n)]＝x²(n)-x(n+1)x(n-1) (17)

for the present invention, the characteristic modal IMF is found_t(n) Teager-Kaiser energy operator TKEO_t(n) the formula is:

TKEO_t(n)＝[IMF_t(n)]²-IMF_t(n+1)IMF_t(n-1) (18)

in the formula (18), IMF_t(n)，IMF_t(n+1)，IMF_tAnd (n-1) is the characteristic mode corresponding to the nth, the (n + 1) th and the (n-1) th sampling points respectively.

Examples

A 10kV radial power distribution network model as shown in fig. 2 is established, and 6 feeders are provided in total, wherein the parameters of the overhead lines and the cable lines are as shown in table 1:

TABLE 1

FIG. 3, FIG. 4, FIG. 5 show the feed line l₆Zero sequence currents when HIF, CS and LS occur respectively, and as can be seen from FIG. 3, when HIF occurs, nonlinear distortion occurs when the current waveform crosses zero, and the waveform as a whole shows a gradually increasing trend; as can be seen from fig. 4, when CS, a high-frequency disturbance component is more likely to be generated in the current; as shown in fig. 5, when the LS switching is unbalanced, the originally balanced distribution network in fig. 2 is broken, and at this time, a zero-sequence current occurs, and the zero-sequence current generates a certain high-frequency component at the instant of the LS switching, and then the current waveform tends to be stable. Therefore, as can be seen from fig. 3, 4 and 5, the zero sequence currents of HIF, CS and LS have certain differences, and the differences can be used to construct the identification criterion.

Table 1 shows the calculation results, and it can be seen that TS of 4 subintervals is used in the case of CS and LS_{t_x}The values are all 0, and then the final total time entropy TS is also 0; TS of 4 subintervals Only when HIF occurs_{t_x}If the value is not 0, the final total time entropy value TS is not 0 after the summation, whereby the decision according to the invention is madeAccording to the recognition result, the same results are obtained for the overhead wire l in Table 1₁Cable line l₂Cable-wire hybrid wire₆。

TABLE 1

Through the mode, the transient zero sequence current is subjected to feature extraction by adopting a variable modal decomposition algorithm with self-adaptive characteristics, and a series of intrinsic modal components are obtained; a kurtosis calculation formula is introduced, a mode corresponding to the maximum kurtosis value is selected to be defined as a characteristic mode, the accurate extraction of transient zero sequence current characteristics is realized, a high-frequency component reflecting sensitive change is obtained, and a foundation is laid for the construction of the next detection criterion. The detection criterion construction aspect is as follows: aiming at the defects of large calculated amount and the like of the existing network identification method, a Teager-Kaiser energy operator calculation characteristic mode is provided, and the Teager-Kaiser energy operator during high-resistance fault, capacitor switching and load switching is obtained; further, a time entropy calculation theory is constructed by carrying out interval division on the Teager-Kaiser energy operator, and total time entropy values of HIF, CS and LS are obtained; finally, by identifying whether the total time entropy is 0, accurate detection of HIF is achieved. The intermittent reignition and extinguishing characteristics of the high-resistance fault current are fully utilized: the intermittent re-ignition and extinguishing phenomena of HIF current waveforms near zero crossing points are fully utilized, the CS and LS switching are instantaneous disturbance, zero sequence current waveforms of the HIF current waveforms show oscillation attenuation trend, a characteristic mode reflecting the law is accurately extracted by adopting a VMD method, and the HIF is accurately identified through Teager-Kaiser energy operator and time entropy calculation.

Claims (7)

1. The fault detection method based on the fault current intermittent reignition and extinguishment characteristics is characterized by being implemented according to the following steps:

step 1, carrying out mode-changing decomposition on zero-sequence current to obtain each intrinsic mode, calculating kurtosis values of each intrinsic mode, and defining the mode corresponding to the maximum kurtosis value as a characteristic mode;

step 2, calculating Teager-Kaiser energy operators for the characteristic modes, and carrying out interval division on the waveforms of the energy operators;

step 3, calculating the time entropy value of each interval to finally obtain the total time entropy value;

and 4, when the total time entropy is not 0, judging that the power distribution network has a high-resistance fault, and when the total time entropy is equal to 0, judging that the power distribution network has capacitor switching or load switching.

2. The fault detection method based on the fault current intermittent reignition and extinguishment characteristic as claimed in claim 1, wherein the specific process of the step 1 is as follows: respectively decomposing the zero sequence current of the power distribution network by adopting a variable modal decomposition algorithm to obtain intrinsic modal components IMF₁(n)，IMF₂(n)，…，IMF_s(n), s is a mode number, and n is a sampling point; for IMF₁(n)，IMF₂(n)，…，IMF_s(n) separately calculating kurtosis values k₁，k₂，…，k_s(ii) a At kurtosis value k₁，k₂，…，k_sTo find out the maximum kurtosis value k_tThen the maximum kurtosis value k_tCorresponding intrinsic mode component IMF_tAnd (n) is defined as a characteristic mode of the zero sequence current of the power distribution network.

3. The fault detection method based on fault current intermittent reignition and extinguishment characteristics as claimed in claim 2, wherein the modal component IMF is_sKurtosis value k of (n)_sThe calculation formula is as follows:

wherein, mu and sigma are intrinsic mode components IMF_sAnd (N) the mean value and the standard deviation, wherein N is the total number of sampling points.

4. The fault current based intermittence of claim 1The fault detection method of the nature reignition and extinguishment characteristics is characterized in that the specific process in the step 2 is as follows: for characteristic mode IMF_t(n) calculating Teager-Kaiser energy operator TKEO_t(n); to energy operator TKEO_tThe waveform of (n) is divided into intervals at intervals, and the interval Δ T of the interval division is 0.01s, thereby obtaining an interval 1: TKEO_{t_1}(n), interval 2: TKEO_{t_2}(n), …, interval x: TKEO_{t_x}(n), x is the interval number.

5. The fault detection method based on fault current intermittent reignition and extinguishment characteristics as claimed in claim 4, wherein IMF is applied to the characteristic mode_t(n) calculating Teager-Kaiser energy operator TKEO_tThe calculation formula of (n) is as follows:

TKEO_t(n)＝[IMF_t(n)]²-IMF_t(n+1)IMF_t(n-1)

in the above formula, IMF_t(n)，IMF_t(n+1)，IMF_tAnd (n-1) is the characteristic mode corresponding to the nth, the (n + 1) th and the (n-1) th sampling points respectively.

6. The fault detection method based on the fault current intermittent reignition and extinguishment characteristic as claimed in claim 1, wherein the specific process of the step 3 is as follows: for each interval TKEO_{t_1}(n)，TKEO_{t_2}(n)，…，TKEO_{t_x}(n) separately calculating the time entropy values TS_{t_1}，TS_{t_2}，…，TS_{t_x}(ii) a For time entropy values TS of intervals_{t_1}，TS_{t_2}，…，TS_{t_x}And adding and summing to obtain a total time entropy value TS.

7. The fault detection method based on intermittent reignition and extinction characteristics of fault current as claimed in claim 6, wherein the interval TKEO_{t_x}(n) the calculation of the time entropy value is as follows:

TS_{t_x}＝-∑p_{t_x}(m)[lnp_{t_x}(m)]

where m is the sampling point in the interval x, p_{t_x}(m) is the interval x energy operator TKEO_{t_x}(n) the probability that the number of peaks accounts for the number of peaks in the total interval;

the total time entropy value TS is calculated as follows:

wherein l is the total number of intervals.

Images