CN105738764B - Fault Section Location of Distribution Network based on transient information Whole frequency band - Google Patents

Abstract

The invention discloses a method for locating a distribution network fault section based on the full frequency band of transient information, which comprises the following steps: collecting fault transient signals of each monitoring point of a distribution line; performing denoising and filtering on the fault transient signals, Obtain the transient filter signal; judge the fault initial phase angle of the transient filter signal of the distribution line; if the fault initial phase angle is zero, use the EMD method to decompose the transient filter signal to obtain the DC component, and use the DC component to analyze the fault and if the initial phase angle of the fault is non-zero, use the adaptive harmonic wavelet packet transform to subdivide the entire frequency band of the transient filter signal to determine the characteristic frequency band, and extract the characteristic frequency band corresponding to each monitoring point The extreme value and polarity of the wavelet transform coefficients are used to locate the fault section. The invention makes full use of the transient full frequency band information at the time of the fault, and accurately judges the fault section according to different initial phase angles of the ground fault.

Description

Fault section location method in distribution network based on full frequency band of transient information

technical field

The invention relates to a method for locating a distribution network fault section based on a full frequency band of transient state information.

Background technique

With the access of distributed energy sources and the development of smart distribution network construction, the existing distribution network fault location technology mainly adopts the line selection or direct protection trip through the protection device in the station or the small current line selection device, and then manually The search method is used to locate and troubleshoot the fault point. Since the distribution network is mainly close to the user side, when the fault point occurs in an area with complex terrain or the fault trace is not obvious, relying on manual line inspection will greatly prolong the power supply recovery time, and the efficiency is low. In line with the development trend of the future distribution network.

At present, the application of traveling wave method in transmission lines has been very mature. However, compared with transmission lines, there are a large number of T-connections in distribution networks, and most of the line ends are users. The double-terminal traveling wave distance measuring device is not suitable for distribution networks.

In recent years, the distribution network has installed fault indicators for fault location. Due to the simple principle and low sampling rate of the fault indicator, it can only be applied to large current faults such as phase-to-phase short circuit and two-phase grounding. Fault indicators often report false alarms during ground faults, and the probability of single-phase ground faults in distribution networks is the highest.

On the basis of realizing small current line selection, it is a focus of current research to further determine the fault section and reduce the scope of inspection.

However, because the single-phase ground fault current is very weak compared with the load current, and the ground fault condition is complex, it is difficult to determine the real fault location based on the traditional power frequency or harmonic component fault location method.

Chinese patent document CN 103454559 B discloses a method for locating a single-phase ground fault section of a distribution network, which includes terminals installed at multiple positions on the line in real time detecting the secondary synthesis of the zero-sequence current of the current transformer at the installation position; when any After the zero-sequence current amplitude detected by the terminal exceeds the preset start-up value, all terminals immediately and accurately capture the zero-sequence current transient signal of 1 cycle before the zero-sequence current exceeds the start-up value and 3 cycles after exceeding the start-up value; each terminal Using Prony algorithm, wavelet packet algorithm, HHT algorithm and fractal algorithm to analyze and calculate the zero-sequence current transient signals of four cycles, extract the zero-sequence current characteristic data of each algorithm, and upload them to the master station; master station After receiving the zero-sequence current characteristic data of each algorithm from each terminal, the number of BP neural network layers and the number of input and output neuron nodes of each layer are determined according to the zero-sequence current characteristic data of each algorithm extracted by each terminal, Determine the network structure and network parameters of the BP neural network.

The above-mentioned BP neural network is a prediction system, which brings convenience to the research, and does not need to provide an accurate correspondence mechanism between multiple zero-sequence current characteristic data and fault sections. However, the BP neural network requires a large number of samples for training , and the problem of inaccurate prediction is difficult to eradicate.

Contents of the invention

The purpose of the present invention is to provide a method for locating a distribution network fault section based on the full frequency band of transient state information, so as to accurately locate the fault section.

Therefore, the present invention provides a distribution network fault section location method based on the full frequency band of transient information, comprising the following steps: collecting fault transient signals at each monitoring point of the distribution line; removing fault transient signals Noise filtering to obtain the transient filtering signal; judge the fault initial phase angle of the transient filtering signal of the distribution line; if the fault initial phase angle is zero, use the EMD method to decompose the transient filtering signal to obtain the DC component, and use the DC component to locate the fault section; and if the initial phase angle of the fault is non-zero, use the adaptive harmonic wavelet packet transform to subdivide the entire frequency band of the transient filter signal to determine the characteristic frequency band, and extract the corresponding The extreme value and polarity of the wavelet transform coefficients in the characteristic frequency band are used to locate the fault section.

The method for locating the fault section of the distribution network further includes the step of: using the transient filter signals on both sides of the minimum fault section to determine the location of the fault point or the fault branch by using the double-terminal traveling wave method.

In the above distribution network fault section positioning method, the fault transient signal of each monitoring point of the distribution line is collected by the measurement terminal, wherein each measurement terminal detects in real time whether the zero-sequence voltage of the distribution line exceeds the limit, if the limit Then the 2 cycles before the fault start time and the 4 cycles after the start time are taken as fault transient signals.

In the above distribution network fault section location method, the RLS adaptive filter is used to denoise and filter the fault transient signal, in which the two cycles before the fault are used as noise signals, and the four cycles after the fault point are used as periodic signals.

In the above distribution network fault section location method, "Using the DC component to locate the fault section" includes: subtracting the DC components of every two adjacent monitoring points, and finding the two phases with the largest DC component difference. If there are adjacent monitoring points, the fault zone is between these two adjacent monitoring points.

In the above distribution network fault section location method, "using adaptive harmonic wavelet packet transform to subdivide the entire frequency band of the transient filter signal to determine the characteristic frequency band" includes the following steps: A) performing J on the transient filter signal layer harmonic wavelet packet transformation; B) decompose each subband of the jth layer into three subbands with generalized harmonic wavelet, and the three subbands are of equal bandwidth. After decomposing, 3×2 ^j subbands are obtained, where j= 0, 1, 1.6, 2, 2.6, 3, ..., J; C) Bring the above sub-bands between the j+1 and j+2 layers of the harmonic wavelet packet transform to form a new harmonic wavelet packet Transformation; D) Calculating the normalized energy of the jth layer, the sth subband energy relative to the sum of all subband energies of the jth layer, wherein, s=1, 2, 3,..., 2J: E) calculating each subband Spectral kurtosis, and F) Determine the characteristic frequency band according to the principle of maximum spectral kurtosis.

Compared with the prior art, the present invention has the beneficial effects of fully utilizing the transient full-band information at the time of the fault, and using the high-frequency transient capacitive current component and the low-frequency attenuated DC component respectively for different ground fault initial phase angles The judgment of the fault section provides an accurate correspondence mechanism between multiple zero-sequence current characteristic data and the fault section, and is suitable for use in combination with the double-terminal traveling wave method.

In addition to the objects, features and advantages described above, the present invention has other objects, features and advantages. Hereinafter, the present invention will be described in further detail with reference to the drawings.

Description of drawings

The accompanying drawings constituting a part of this application are used to provide a further understanding of the present invention, and the schematic embodiments and descriptions of the present invention are used to explain the present invention, and do not constitute an improper limitation of the present invention. In the attached picture:

Fig. 1 is the flow chart of the distribution network fault section location method based on transient state information full frequency band according to the preferred embodiment of the present invention;

Fig. 2 is a schematic diagram of a distribution network fault section location method according to the present invention being used in a distribution network;

Fig. 3 is a functional block diagram of the RLS adaptive filter applied in the distribution network fault section location method according to the present invention; and

Fig. 4 is a processing method in the case that the fault initial phase angle is non-zero in the distribution network fault section location method according to the present invention.

Detailed ways

It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and examples.

The present invention provides a distribution network fault section location method based on the full frequency band of transient information, as shown in Figure 1 and Figure 2, the method includes the following steps: collecting the fault transient signal of each monitoring point of the distribution line Step S10; Step S20 of denoising and filtering the fault transient signal to obtain a transient filtered signal; Step S30 of judging the fault initial phase angle of the distribution line; if the fault initial phase angle is not Zero, using the adaptive harmonic wavelet packet transform to subdivide the transient filter signal in the full frequency band, and determine the characteristic frequency band, then extract the extreme value and polarity of the wavelet transform coefficients in the characteristic frequency band of each monitoring point, and correct the fault Step S40 of locating the section; if the initial phase angle of the fault is zero, decompose the transient filter signal by EMD method to obtain a DC component, and use the DC component to locate the fault section S50; and Step S60 using the terminal filtering data on both sides of the minimum fault section and using the double-ended traveling wave method Db wavelet transform to calculate the head time of the transient traveling wave to determine the location of the fault point or fault branch.

Compared with the prior art, the present invention has the beneficial effects of fully utilizing the transient full-band information at the time of the fault, and using the high-frequency transient capacitive current component and the low-frequency attenuated DC component respectively for different ground fault initial phase angles It can judge the fault section and is suitable for use in combination with the double-ended traveling wave method.

The present invention utilizes feature quantity (direct current component, extremum and polarity of the wavelet transform coefficient in the characteristic frequency band of each monitoring point) to carry out fault section location, and specific criterion is as follows:

When a small current fault is grounded, the ground transient current can be expressed by the following general formula (1):

Among them, the first item on the right side of the equal sign is the steady-state component of the ground current, which is equal to the difference between the steady-state capacitive current and the inductive current, and the rest are transient components of the ground current, which are equal to the difference between the transient component of the capacitive current and the transient DC component of the inductor current. and.

It can be seen from the above formula that when the ground fault occurs at the maximum phase voltage, that is, π/2, the transient current is mainly transient capacitive current, and the inductive DC component is close to 0. When the ground fault occurs near the zero-crossing point of the phase voltage, The transient current is mainly the inductive DC component and is close to the maximum value, and the capacitive current is close to zero. Since most of the faults occur when the phase voltage is at the maximum value, the small probability occurs when the phase voltage is zero.

According to the above fault characteristics, if the monitored line fault occurs when the phase voltage is maximum, the fault section location can be carried out according to the extreme value and polarity of each wavelet transform coefficient/mode in the characteristic frequency band; is zero, then the DC component obtained by decomposing the transient filter signal according to the EMD method is used to locate the fault section.

The present invention provides a better solution for implementing the above step S10, which is as follows: install a distributed high-density sampling terminal on the monitored distribution line for collecting the three-phase voltage and current signals of the outdoor overhead line, and through Karenbauer transformation, The transient filtering data can be decomposed into zero-mode components and line-mode components, and the zero-sequence current and zero-sequence voltage components are calculated by the following formulas (2) and (3):

Real-time detection of whether the zero-sequence voltage U ₀ exceeds the limit, and if it exceeds the limit, the transient filter will be started. The zero-sequence voltage limit threshold is (0.15 times) the rated bus voltage; 4 cycles after time.

The numbers ①～⑩ on the line are the 10 measurement terminals installed on the line. As shown in Figure 2, the three-phase voltage and three-phase current are obtained respectively, and each two terminals realize the division of the faulty section of the faulty line. Among them, the terminal 3, 4, 6, 7, 9, and 10 are installed at the end of the line to capture fault transient voltage components, and the rest capture fault transient current components.

The present invention provides a better solution for implementing the above step S20 to solve the problem that various random noises and unbalanced currents and other interference factors often exist in the three-phase current during the operation of the power distribution system. When the ground fault resistance is relatively large, the fault transient The problem of low component signal-to-noise ratio, the solution is as follows:

The RLS adaptive filtering algorithm is used to filter the collected signals, as shown in Figure 3, where the two cycles before the fault are taken as the noise signal d(k), and the four cycles after the fault point are taken as the periodic signal x(k). The filter is designed with an 8th-order transverse FIR filter and implemented with a FPGA. After filtering, the output signal is y(k), and e(k) is the error signal.

The present invention provides the better scheme that implements above-mentioned step S40 to the fault zero-sequence current transient signal of collection to carry out self-adaptive harmonic wavelet packet transformation, as shown in Figure 4, specifically comprises the following steps:

S41) carry out J=5 layers of harmonic wavelet packet transformation to signal, wavelet basis function adopts Db8 wavelet function;

S42) Decompose each subband of the jth layer into three subbands with generalized harmonic wavelet, and the three subbands are of equal bandwidth. After the above decomposition, 3×2 ^j subbands are obtained;

S43) These sub-bands are brought between the j+1 layer and the j+2 layer of the harmonic wavelet packet transform to form a new harmonic wavelet packet transform;

S44) Calculate the normalized energy γ(j, s) of the j-th layer and the s-th sub-band energy relative to the sum of all sub-band energies of the j-th layer according to the following formula (4):

in is the wavelet coefficient of the jth layer and the sth subband, and N is the length of each subband coefficient;

S45) According to the fault transient component characteristics, a threshold λ is set, and the spectral kurtosis of each subband is calculated according to formula (5):

in, Spectral kurtosis estimated for the jth layer and the sth subband;

S46) According to the principle of maximum spectral kurtosis, determine the optimal frequency band FS as the characteristic frequency band; and

S47) After the characteristic frequency band is determined, extract the modulus maximum value and polarity of the wavelet transform coefficients in the characteristic frequency band.

The present invention provides the preferred scheme of implementing step S50, specifically as follows: Utilize the EMD method to decompose the transient filter signal, calculate and obtain the IMF component and the residual component under each frequency according to the following formula (6):

In the above formula, r _n (t) is the DC component of transient signal decomposition;

The present invention provides a better solution for implementing step S60, which is as follows: use the terminal filtering data on both sides of the minimum fault section, use the double-terminal traveling wave method and Db8 wavelet transform, calculate the head time of the transient traveling wave, and further determine the fault point position or fault branch, the double-ended traveling wave calculation formula (7) is as follows:

Where l _f is the distance from the measurement point, l is the distance between two measurement terminals, v ₁ is the propagation speed of the traveling wave, and t ₂ and t ₁ are the moments when the traveling wave reaches both ends.

Below in conjunction with Fig. 1 and Fig. 2, introduce the section positioning method of the present invention in detail:

Assuming that the fault point is located at the arrow between terminals 5 and 7, the initial phase angle of the fault is non-zero, and the zero-sequence currents of terminals 1, 2, 5, 8, and 9 on the fault line are processed by adaptive harmonic wavelet packet transformation to determine Optimal fault frequency band, and calculate the polarity of the transient component after wavelet transform as shown in Table 1:

Terminal number ① ② ③ ⑧ ⑨ polarity + + + — —

It can be seen from the above table that the fault point is located on the line between terminal 2 and terminal 5. The transient signal of terminal 2 and terminal 5 is transformed by Db8 wavelet, and the fault point is calculated according to the double-ended traveling wave formula. If the fault The point is just at the branch point where the lines of terminals 5, 7 and terminals 5, 8 meet. It shows that the fault point is located on the fault branch. The measurement terminal installed at the end of the branch can be used to continue the double-terminal traveling wave location, and finally find the fault point.

The above positioning method is aimed at the case where the initial phase angle of the fault is non-zero, that is, the transient component of the capacitive zero-sequence current is relatively obvious. When the fault occurs at the moment when the initial phase angle is zero, the section can be located according to the DC component decomposed by EMD. That is, carry out EMD decomposition on terminals 1, 2, 5, 8, and 9, extract the attenuated DC component r _n (t), subtract each adjacent two terminals, and find the two adjacent terminals with the largest DC component difference, then It can be determined that the line segment between the two terminals is the faulty road segment.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (6)

1. A distribution network fault section location method based on the full frequency band of transient information, comprising the following steps: (1) Acquisition of fault transient signals at each monitoring point of the distribution line; (2) carry out denoising filter to described fault transient signal, obtain transient filter signal, it is characterized in that, also comprise the following steps: (3) judge the fault initial phase angle of the distribution line; (4) If the fault initial phase angle is zero, utilize the EMD method to decompose the transient filter signal to obtain a DC component, and use the DC component to locate the fault section; and (5) If the fault initial phase angle is non-zero, use adaptive harmonic wavelet packet transform to carry out full-band subdivision of the transient filter signal to determine the characteristic frequency band, and extract the characteristic frequency band corresponding to each monitoring point The extreme value and polarity of the inner wavelet transform coefficients are used to locate the fault section. 2. distribution network fault section localization method according to claim 1, is characterized in that, also comprises the step after described step (5): (6) Using the transient filtered signals on both sides of the minimum fault section, the double-ended traveling wave method is used to determine the location of the fault point or the fault branch. 3. distribution network fault section localization method according to claim 1, is characterized in that, in step (1), utilizes the fault transient signal of each monitoring point of distribution line to collect distribution line, wherein, each institute The measurement terminal detects in real time whether the zero-sequence voltage of the distribution line exceeds the limit, and if it exceeds the limit, the 2 cycles before the fault start time and 4 cycles after the start time are used as fault transient signals. 4. The distribution network fault section location method according to claim 3, characterized in that, in step (2), the fault transient signal is denoised and filtered using an RLS adaptive filter, wherein the fault The first two cycles are used as noise signals, and the four cycles after the fault point are used as periodic signals. 5. The distribution network fault section location method according to claim 1, characterized in that, in step (4), "using the DC component to locate the fault section" comprises: placing every adjacent two The DC components of the monitoring points are subtracted, and two adjacent monitoring points with the largest difference in DC components are found, then the fault section is between the two adjacent monitoring points. 6. The distribution network fault section location method according to claim 1, characterized in that, in step (5), "use the adaptive harmonic wavelet packet transform to carry out full-band fine-tuning of the transient filter signal points to determine the characteristic frequency band" includes the following steps: A) performing J-level harmonic wavelet packet transformation on the transient filter signal; B) Decompose each subband of the jth layer into three subbands with generalized harmonic wavelet, and the three subbands are of equal bandwidth. After decomposing, 3×2 ^j subbands are obtained, where j=0,1,1.6,2 ,2.6,3,...,J; C) above-mentioned sub-band is substituted between the j+1 layer and the j+2 layer of harmonic wavelet packet transform, forms new harmonic wavelet packet transform; D) Calculating the normalized energy of the jth layer, the sth subband energy relative to the sum of all subband energies of the jth layer, wherein, s=1,2,3,..., ^2J ; E) Calculate the spectral kurtosis of each subband, and F) Determine the characteristic frequency band according to the principle of maximum spectral kurtosis.