CN113447761A

CN113447761A - Power distribution network small current grounding fault section positioning method based on maximum mean difference

Info

Publication number: CN113447761A
Application number: CN202110796002.6A
Authority: CN
Inventors: 尹项根; 谭力铭; 王义凯; 乔健; 徐雯; 陈卫; 张哲�; 文明浩
Original assignee: Huazhong University of Science and Technology
Current assignee: Huazhong University of Science and Technology
Priority date: 2021-07-14
Filing date: 2021-07-14
Publication date: 2021-09-28
Anticipated expiration: 2041-07-14
Also published as: CN113447761B

Abstract

The invention discloses a power distribution network low-current ground fault section positioning method and system based on maximum mean difference, and belongs to the field of power distribution network fault positioning. The method comprises the following steps: when the system has single-phase earth fault, determining a fault line; calculating the Maximum Mean Difference (MMD) value between double-end zero-mode currents of each section of the fault line, and taking the section with the maximum MMD as a fault section; the double-end zero-mode current of each section is acquired at the same time. The method is suitable for the problem of positioning the fault section of the power distribution network with the neutral point small current grounding, can effectively solve the problem of positioning section errors caused by zero-mode current similarity characteristic distortion existing in the positioning blind area in the power distribution network line, can accurately position the fault section in different fault scenes, has certain robustness, and can be used as an important component of a power distribution automation system.

Description

Power distribution network small current grounding fault section positioning method based on maximum mean difference

Technical Field

The invention belongs to the field of power distribution network fault location, and particularly relates to a method and a system for locating a power distribution network low-current ground fault section with a maximum mean value difference.

Background

A method for rapidly and accurately positioning fault sections of a power distribution network with frequent single-phase earth faults is a basis for realizing distribution network automation. In general, after a single-phase ground fault occurs, the double-end zero-mode current waveforms of the fault section are low in similarity degree and opposite in polarity, and the double-end zero-mode current waveforms of the non-fault section are high in similarity degree and same in polarity. However, due to the complex structure of the power distribution network and the common overhead-cable hybrid line, when a single-phase ground fault occurs at some position of the line, the zero-mode current waveform similarity degree characteristic may be distorted, and the zero-mode current waveform similarity degree of the fault section is higher than that of the non-fault section, which seriously affects the accuracy of the existing fault section positioning result based on the zero-mode current similarity characteristic, and these special fault positions are called as positioning blind areas.

The existing power distribution network section positioning method is mostly based on zero-mode current characteristics of fault lines, including waveform similarity and polarity relationship. Description methods according to the similarity of zero-mode current waveforms can be classified into a relative entropy method, a transient gravity center frequency method and the like; the description mode according to the polarity direction of the zero-mode current can be divided into a correlation coefficient method, a fault direction measuring method and the like. However, the above methods all have certain limitations, and the former ignores polarity information, and is easy to cause misjudgment of a fault section when a fault occurs in a positioning blind area; the latter only uses the polarity relationship, is seriously influenced by the TA reverse connection problem which often occurs on site, and simultaneously, the situation of undefined polarity can occur due to the short duration of the zero-mode current and the small signal intensity. In recent years, some methods use a combination of the similarity degree and polarity relationship of zero-mode current waveforms. For example, Liupenghe et al, in the power grid technology 2016,40(03): 952-.

With the application of advanced measurement systems such as a power distribution network synchronous phasor measurement unit (D-PMU) and the like in a power distribution network, synchronous measurement of section double-end zero-mode current can be completely realized, and the requirement on the synchronous error resistance of the method is reduced. Considering that the zero-mode current polarity characteristic is unchanged when the positioning blind area has a ground fault, and the reliability of the method based on the zero-mode current polarity characteristic is poor, a method capable of sensitively reflecting the polarity relationship and comprehensively reflecting the waveform similarity degree needs to be found to solve the problem of the positioning blind area.

Disclosure of Invention

Aiming at the defects or improvement requirements of the prior art, the invention provides a method and a system for positioning a small current grounding fault section of a power distribution network with the maximum mean value difference, and aims to solve the problem that a positioning section is wrong due to zero-mode current similarity characteristic distortion existing in a positioning blind area in a distribution network line, correctly position the fault section and improve the power supply reliability.

In order to achieve the above object, the present invention provides a method for positioning a low current ground fault section of a power distribution network based on a maximum mean difference, comprising:

s1, when a single-phase earth fault occurs in a system, determining a fault line;

s2, calculating an MMD value between double-end zero-mode currents of each section of the fault line, and taking the section with the maximum MMD as a fault section; the double-end zero-mode current of each section is acquired at the same time.

Further, the double-end zero-mode current of each section is measured by a synchronous measuring device.

Further, the synchronous measurement device is a PMU.

Further, each section of the fault line is divided according to the line structure and the distribution position of the synchronous measuring device along the line.

Further, the MMD value between the two-terminal zero-mode currents of each section is calculated by using the following formula:

k (x, y) is a kernel function to be selected, distribution p and q respectively represent double-end zero-mode current of each section, sample space x and y represent sampling values corresponding to the zero-mode current, and the maximum mean difference of each section on the fault line is calculated through the formula.

Further, the process of determining the faulty line specifically includes:

s2.1, calculating to obtain bus zero sequence voltage according to the bus three-phase voltage obtained by monitoring, and judging that the system has a single-phase earth fault when the bus zero sequence voltage is more than N times of the bus rated voltage; n is a set constant;

and S2.2, comparing the transient zero-mode currents of the lines on the bus obtained through monitoring, and judging the line with the maximum transient zero-mode current as a fault line.

In general, the above technical solutions contemplated by the present invention can achieve the following advantageous effects compared to the prior art.

The invention provides a power distribution network low current ground fault positioning method based on maximum mean difference, and aims to solve the problem of positioning blind areas existing in power distribution network section positioning. The method judges the fault section according to the calculated value of the maximum mean difference, enables the maximum mean difference to comprehensively reflect the similarity and polarity relationship between the zero-mode currents by selecting proper kernel function types and parameters, is more sensitive to the polarity relationship compared with a dynamic bending distance method by verification, can effectively solve the problem of positioning blind areas caused by similarity characteristic distortion of the zero-mode currents, can accurately position the fault section in different fault scenes, has certain robustness, accelerates the efficiency of fault processing and power supply recovery, improves the power supply reliability of a power distribution network, and can be used as an important component of a power distribution automation system.

Drawings

FIG. 1 is a flow chart of a method for a power distribution network low current ground fault section based on maximum mean difference according to the present invention;

figure 2 is a schematic diagram of a simple distribution line fault provided by the present example;

fig. 3 is a schematic diagram of a fault in the hybrid distribution line provided in the present example;

FIG. 4 is a simplified zero modulus equivalent circuit schematic provided by the present invention;

FIG. 5 is a schematic diagram comparing MMD and DTW calculated values provided by the present invention;

fig. 6 is a schematic diagram of the zero mode current distortion of the dead zone of the localization provided by the present example.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

The method for positioning the small-current ground fault section of the power distribution network based on the maximum mean difference, provided by the embodiment of the invention, has the flow shown in fig. 1, and comprises the following steps:

specifically, the invention does not limit the fault line selection method; alternatively, the faulty line determination is called:

calculating to obtain bus zero sequence voltage according to the bus three-phase voltage obtained by monitoring, and judging that the system has a single-phase earth fault when the bus zero sequence voltage is greater than 0.15 times of the bus rated voltage; and comparing the transient zero-mode currents of all the lines on the bus obtained by monitoring, and judging the line with the maximum transient zero-mode current as a fault line.

In an embodiment of the invention, the line structure and the synchronous measurementThe installation location of the device is shown in the simple distribution line fault model of fig. 2 and the hybrid distribution line fault model of fig. 3. In the circuit f, as illustrated in FIG. 2₁When a point has a fault, transient zero-mode current waveform data of the

measuring points

1, 2, 3, 4, 5 and 6 after the fault needs to be acquired, double-end zero-mode currents corresponding to each section are grouped, and the zero-mode current corresponding to the section I is the measuring point 1 and the measuring point 2; the zero-mode current corresponding to the section II is a measuring point 3 and a measuring point 4; the zero mode current for segment III is measured at points 5 and 6. The MMD values for each segment are calculated in three groups. Calculating MMD values among all groups of zero-mode currents, namely MMD values among two-terminal zero-mode currents of corresponding sections;

the MMD method is used for comparing the difference between different distributions, projecting sample spaces of different distributions from an original space to a feature space through a function f, and respectively calculating the mean value, and expressing the mean value difference as the difference between the distributions. When the function F is expanded into a function set F, the maximum value of the mean difference is called the maximum mean difference, and its expression is as follows:

wherein: wherein sup represents the supremum of the mean difference; p and q represent two different distributions; x and y represent the sample space corresponding to the distribution; e_x～p[f(x)]Which represents the mean value after the sample space x has been projected into the feature space by the function f.

The MMD is closely related to the selection of the function set F, and in order to ensure the diversity and convergence of the function set F, relevant documents prove that the function set F meets relevant requirements when mapping a sample space to a unit sphere in a regenerated Hilbert space.

Defining R as original space, H as characteristic space, and mapping function as

The relationship between the kernel function and the mapping function is

In the present inventionIn the light background, it is desirable that the kernel function can reflect the similarity degree and polarity relation of zero-mode current, and the radial basis function satisfies the requirement. When k (x, y) ═ k (| | x-y |), if the polarity of the zero-mode current is opposite, the euclidean distance | | | | x-y | | | between x and y is larger than that when the polarity of x and y is the same, and the polarity relationship can be better reflected, so that the radial basis function is selected as the kernel function, and the commonly used gaussian kernel function can be selected as follows:

the kernel coefficient σ is affected by the magnitude of the sampled values x and y, and the zero-mode current of the system is usually between several amperes and several tens of amperes, so in the context of the method of the present invention, the kernel coefficient may be set to be σ 10. The maximum mean difference calculated by the kernel function is expressed as follows:

wherein, distribution p and distribution q represent the two-terminal zero-mode current distribution of the segment respectively; sample value x_iAnd y_jRespectively representing ith or jth zero-mode current sampling values at two ends of the section; and m and n respectively represent the number of zero mode current sampling values at two ends of the section. And calculating the maximum mean difference of each section on the fault line through the formula.

In order to verify that the MMD method can better reflect the polar relationship compared with the DTW method, the zero-mode current is simplified into the sinusoidal current with single characteristic frequency, and the attenuation process is ignored. The simplified fault line zero-modulus equivalent circuit is shown in FIG. 4: assume a zero-mode current of i at M0_M01.5sin (2000 π t), another position is defined as zero-mode current i_x＝Ai_M0. According to the distribution rule of the zero-mode current, the directions of the zero-mode current at the upstream and the downstream of the fault point are opposite, and the closer to the fault point, the larger the zero-mode current is. Therefore, if 0<A<1，i_xCan be represented as i_M1And i_M0Double-end zero-mode current forming a normal section II; if A<0，i_xCan be represented as i_N0And i_N0Double-ended zero mode current constituting fault section III. Firstly, the relation that the MMD and the DTW between the section III and the section II change along with the parameter A is discussed, the sampling rate is assumed to be 10kHz, and the sampling data of one power frequency period after each point fault is taken for comparison. The variation relationship between MMD and DTW with respect to parameter A is shown in FIG. 5, the left end A of which<0, corresponding to a fault section III; right end 0<A<1, corresponding to normal segment II. The MMD calculation value of the sector III must be larger than the sector II, and the DTW calculation value of the sector III may be smaller than the sector II, so that the DTW method will misjudge the sector II as the fault sector, and the MMD method correctly judges the sector III as the fault sector. As can be seen from fig. 5, the MMD method is more sensitive to the polarity relationship between the signals than the DTW method, and the MMD value is relatively larger when the two signal polarity relationships are opposite. It should be noted that fig. 5 only verifies that the method of the present invention is more sensitive to the polarity relationship than the DTW method, and does not represent the actual calculated value.

According to the embodiment of the invention, a simple distribution network line fault model shown in figure 2 and a hybrid distribution line fault model shown in figure 3 are built in a PSCAD/EMTDC software platform. The same frequency change parameters are respectively adopted by the overhead line and the cable line, the sampling frequency is set to be 10kHz, 50 groups of transient zero-mode current data of 5ms after the fault are collected and used for calculating the MMD value, the positioning effect of the power distribution network small current grounding fault section positioning method based on the maximum mean difference is compared, and meanwhile, the DTW value is calculated to perform fault section positioning as comparative analysis.

In FIG. 2, failure point f₁Fault point f₂And point of failure f₃The fault line main part, the fault line branch part and the positioning blind area are respectively arranged. When f is₃When a single-phase earth fault occurs, the waveforms of the measuring point 7, the measuring point 8 and the measuring point 9 are shown as a zero-mode current distortion schematic diagram of a positioning blind area in fig. 6. The absolute values of the correlation coefficients of the zero mode current waveforms corresponding to the measuring points 7 and 8, and the measuring points 8 and 9 (i.e., the sections IV and V) are calculated as shown in table 1:

TABLE 1 Absolute values of correlation coefficients for double-ended zero-mode currents for section IV and section V

The closer the absolute value of the correlation coefficient is to 1, the higher the waveform similarity of the correlation coefficient is, and it can be seen from table 1 that when a fault occurs at some position of the distribution line, the higher the similarity of the waveform of the double-end zero-mode current of the fault section (section IV) is than that of the non-fault section (section V), the distortion of the zero-mode current similarity characteristic may occur, and at this time, the non-fault section may be misjudged as the fault section by the existing section positioning method.

In order to facilitate the demonstration of the section positioning result of the method of the present invention, the MMD value and the DTW value of each section are normalized, the MMD value and the DTW value of the section corresponding to the section with the maximum MMD value and DTW value are set to 1, the MMD values and the DTW values of other sections are scaled in equal proportion, the section with the distance between the MMD value and the DTW value of 1 is defined as a fault section, and the calculation result is shown in table 2.

TABLE 2 simple overhead line distribution network line fault section location results

As can be seen from Table 2, when the system trunk line f₁And branch line f₂When single-phase earth faults occur, the MMD method and the DTW method can accurately judge fault sections. And when the dead zone f is positioned₃When a single-phase earth fault occurs, the DTW method misjudges the section V as a fault section, and the MMD correctly judges the fault section as a section IV.

In FIG. 3, the center points f of the sections I are respectively taken into consideration₁And mid-point f of section IV₂The positioning results of the MMD method and the DTW method are calculated when a metallic single-phase ground fault occurs, as shown in table 3.

TABLE 3 results of segment location at mixed line fault

As can be seen from the positioning results shown in table 3, the MMD method for determining the single-phase ground fault at any position in the overhead line-cable hybrid line can accurately determine the fault position, and the DTW method may cause erroneous determination.

Further considering the influences of different neutral point grounding modes, fault phases, fault angles and transition resistances, the fault location scenarios and fault location results based on the MMD method in the simple overhead line and the hybrid line under different fault conditions are respectively shown in tables 4 and 5.

Scene numbers

1, 2, 3, 4, 5, and 6 correspond to the simple distribution line scene shown in fig. 2, and

scene numbers

7, 8, 9, 10, 11, and 12 correspond to the hybrid distribution line fault shown in fig. 3.

TABLE 4 Fault conditions for different scenarios

TABLE 5 Fault location results corresponding to different scenarios

Because the MMD method can comprehensively reflect the waveform similarity degree and the polarity relation between the zero-mode current signals, and is more sensitive to the polarity relation compared with the DTW method, the problem of positioning blind areas in distribution network lines can be effectively solved, and fault sections can be accurately determined in other different fault scenes.

To verify the robustness of the method proposed herein, MMD method-based localization results were calculated for the fault scenarios shown in table 4, taking into account the effects of arc grounding and measurement noise on the localization of the fault section. Noise with a signal-to-noise ratio of 10dB is added to each acquired zero-mode current signal. And meanwhile, a Cassie arc model is built in the PSCAD/EMTDC and parameters are set. The Cassie dynamic arc equation is:

in the formula: g is arc conductance; θ is the time constant of the arc; e is the arc column voltage gradient; e₀The arc voltage gradient at rest is constant on the assumption of Cassie. After adding the arc model and considering the noise with a signal-to-noise ratio of 10dB, the simulation models shown in fig. 2 and 3 and the setting of the scene parameters in table 4 were used, and the results are shown in table 6:

table 6 MMD method fault section positioning result considering robustness

According to the data in the table, it can be seen that the MMD method can still accurately realize the zone positioning and has stronger robustness under the condition of considering the arc grounding model and the measurement noise.

The method can effectively solve the problem of the blind zone of the section positioning, can accurately position the fault section in different fault scenes of the distribution network line through simulation verification, has stronger robustness, effectively improves the power supply reliability, and provides a foundation for fault location and power distribution automation.

It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims

1. A distribution network low current ground fault section positioning method based on maximum mean difference is characterized by comprising the following steps:

s2, calculating the maximum mean value difference MMD value between double-end zero-mode currents of each section of the fault line, and taking the section with the maximum MMD as a fault section; the double-end zero-mode current of each section is acquired at the same time.

2. The method for locating the small-current ground fault section of the power distribution network based on the maximum mean difference as recited in claim 1, wherein the double-end zero-mode current of each section is measured by a synchronous measuring device.

3. The method for locating the small-current ground fault section of the power distribution network based on the maximum mean difference as claimed in claim 2, wherein the synchronous measurement device is a PMU.

4. The method for locating the small-current ground fault section of the power distribution network based on the maximum mean difference as claimed in claim 2, wherein each section of the fault line is divided according to the line structure and the distribution positions of the synchronous measuring devices along the line.

5. The method for locating the small current ground fault section of the power distribution network based on the maximum mean difference as claimed in any one of claims 1 to 4, wherein the MMD value between the two-end zero-mode currents of each section is calculated by using the following formula:

6. The method for locating the distribution network low-current ground fault section based on the maximum mean difference as claimed in any one of claims 1 to 5, wherein the process of determining the fault line is specifically as follows:

7. A distribution network undercurrent ground fault section positioning system based on maximum mean difference, characterized by includes: a computer-readable storage medium and a processor;

the computer-readable storage medium is used for storing executable instructions;

the processor is used for reading executable instructions stored in the computer-readable storage medium and executing the power distribution network small current ground fault section positioning method based on the maximum mean difference as set forth in any one of claims 1 to 6.