Power distribution network section positioning method and device based on fault phase current multi-dimensional feature clustering

Abstract

The invention provides a power distribution network section positioning method based on fault phase current multi-dimensional feature clustering, which comprises the following steps of: when a single-phase earth fault occurs, collecting three-phase current recording data of the head end of a fault line according to a line selection result of a master station, subtracting current of a cycle wave before the fault from current of a cycle wave after the fault, and obtaining a three-phase current recording data of the head end of the fault line; removing a load current, and constructing a three-phase transient current; processing the three-phase transient current by using a box counting dimension method, identifying a fault phase, and uploading a phase selection result to a master station; the main station collects fault phase current recording data of detection points of all sections of the fault line according to a phase selection result, and seven-dimensional time domain features are extracted; fCM clustering is carried out based on the seven-dimensional time domain features, and a fault occurrence position is determined according to a clustering result in combination with a line topological structure.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network fault identification, and in particular to a distribution network section positioning method and device based on multi-dimensional feature clustering of fault phase current. Background Art

[0002] The corridor environment of medium-voltage power distribution lines in cities in my country is complex. Some overhead lines have insufficient wind deflection distance and are easily in contact with medium such as tree branches and dry cement. Some cable lines are in poor operating conditions. Mechanical damage combined with high temperature and high humidity environment can easily accelerate insulation damage, leading to frequent single-phase grounding faults, of which high-resistance grounding faults account for about 10%. The fault current in the early stage of high-resistance grounding is weak, which can easily cause the line overcurrent protection to refuse to operate. The fault section cannot be isolated in time, and the line protection can only be triggered after the grounding fault point is gradually burned and carbonized and transformed into low-resistance grounding, or manual line inspections are arranged section by section. In this process, the grounding arc is difficult to extinguish by itself, and the probability of inducing electrical fires or electric shocks increases sharply.

[0003] At present, the methods for locating high-resistance grounding fault sections in distribution networks are mainly divided into two categories: active locating method and passive locating method. The active locating method injects characteristic signals such as pulse signals and high-frequency signals into the system to capture the flow difference of characteristic signals in the fault section and the sound section to achieve positioning. This method avoids the disadvantage that the original signal characteristics of high-resistance grounding faults are not obvious, and can independently construct fault characteristics according to the line topology structure, but it requires the additional installation of signal generators such as active inverters, which is not economical, and the positioning effect depends on the repeated debugging of the injected signal amplitude and frequency.

[0004] The passive positioning method directly extracts the signal feature differences of each section for positioning, which is convenient to implement and has a wider application. Common methods include positioning by comparing the difference in transient zero-sequence current skewness coefficients of adjacent sections; or by comparing the Hausdorff distances of each section on the basis of the power frequency component of the zero-sequence current to identify the fault section; or by using signal processing methods such as pole symmetric modal decomposition, synchronous squeeze S transform, and wavelet packet transform to extract the transient energy function, low-frequency energy, time-frequency matrix and other characteristic quantities of zero-sequence current in each section for positioning.

[0005] However, there are still many inconveniences in extracting zero-sequence current in each section on site. Some feeders are only equipped with zero-sequence current transformers at the head end of the line, and the downstream lines do not have the conditions for zero-sequence current measurement; the zero-sequence current of some lines is synthesized by three independent current transformers, which can easily lead to zero-sequence current distortion when sampling is not synchronized and transmission characteristics are inconsistent; and the weak fault current of high-resistance grounding is much smaller than the rated value of the CT primary current, resulting in a measurement error higher than the nominal value. In the above cases, the section positioning method that relies on zero-sequence current may cause mispositioning.

[0006] In order to solve the above problems, people have been seeking an ideal technical solution. Summary of the invention

[0007] Based on this, it is necessary to provide a distribution network section positioning method and device that can solve the problem of line overcurrent protection refusing to operate when a high-resistance grounding fault occurs in a low-resistance grounding system, and the failure of traditional positioning methods caused by the difficulty in accurately obtaining zero-sequence current, in response to the above-mentioned technical problems.

[0008] In order to achieve the above object, the present invention provides a first aspect of a distribution network section location method based on multi-dimensional feature clustering of fault phase current, comprising the following steps: When a single-phase grounding fault occurs, the three-phase current recording data at the head end of the fault line is collected according to the line selection result of the master station, and the current of one cycle after the fault is subtracted from the current of one cycle before the fault to eliminate the load current and construct the three-phase transient current; The three-phase transient current is processed using the box counting dimension method to identify the fault phase, and the phase selection result is uploaded to the main station; Based on the phase selection results, the master station collects the fault phase current recording data of the detection points in each section of the fault line and extracts the seven-dimensional time domain features; FCM clustering is performed based on seven-dimensional time domain features, and the fault location is determined based on the clustering results and combined with the line topology.

[0009] In a possible implementation manner, the three-phase transient current is processed by using the box counting dimension method to identify the fault phase, including the following steps: Binarization processing is performed on the three-phase transient current waveform curve at the head end; The grid size is dynamically adjusted in an exponential decreasing manner, and the matrix order corresponding to the binary image is adjusted to a power of 2; From the maximum grid size Starting from, divide the binary matrix P into Sub-matrices, count the number of non-all-1 matrices ; then order ,statistics , and so on, until the minimum grid size is reached ; Last pair and The constructed data series is fitted by the least squares straight line, and its slope is calculated as the box-counting dimension D. The phase with the smallest box-counting dimension D of the phase transient current waveform is the fault phase.

[0010] In a possible implementation, the seven-dimensional time domain features include: the dynamic time bending distance between the fault phase current at each detection point and the fault phase current waveform at the head end of the line, and the root mean square, peak-to-peak value, peak factor, pulse factor, margin factor, and number of zero crossings of the transient current of the fault phase in each section of the fault line.

[0011] In a possible implementation, performing FCM clustering based on seven-dimensional time domain features includes: Setting the objective function J for:

[0012]

[0013] Among them, 2 is the number of cluster centers, including two cluster centers of the fault section and the sound section; γ ik is the membership degree, which is used to characterize the i The probability that a detection point belongs to a faulty section or a sound section; n is the total number of current detection points on the fault line; m is the weight parameter used to control γ ik The degree of blurring; y i Indicates i The transient current characteristic vector of the fault phase at each detection point; μ k Indicates k The feature vector corresponding to the cluster center has the same dimension as x i consistent;|| y i - μ k || represents the Euclidean distance, which is used to measure the feature vector y i and μ k The smaller the Euclidean distance, the closer the i The closer the detection point is to the k Cluster centers; Use the FCM algorithm to find the objective function through multiple iterations J The minimum value of μ k and membership γ ik To perform synchronous update, the formula is as follows:

[0014]

[0015] After updating, continue to calculate the objective function value until convergence to obtain the clustering result.

[0016] In a possible implementation, the steps of obtaining the dynamic time warping distance are: Set the fault phase current sequence at the head end of the fault line P ={ p 1, p 2,…… p n}, fault phase current sequence of healthy section Q ={ q 1, q 2,…… q n}; Constructing cumulative distance matrix O ,in:

[0017]

[0018] Cumulative distance matrix O Filling is performed, where:

[0019] Finally, the two phase current sequences P and Q The DTW distance is O ( n , n ).

[0020] In order to achieve the above-mentioned object, the second aspect of the present invention provides a distribution network section positioning device based on multi-dimensional feature clustering of fault phase current, comprising: The three-phase current acquisition module is used to collect the three-phase current recording data at the head end of the fault line according to the line selection result of the master station when a single-phase grounding fault occurs, and the current of one cycle after the fault is subtracted from the current of one cycle before the fault to eliminate the load current and construct the three-phase transient current; The fault phase screening module is used to process the three-phase transient current using the box counting dimension method, identify the fault phase, and upload the phase selection result to the main station; The feature extraction module is used by the master station to collect the fault phase current recording data of the detection points in each section of the fault line according to the phase selection result, and extract the seven-dimensional time domain features; Clustering module, used for FCM clustering based on seven-dimensional time domain features; The fault determination module is used to determine the fault location based on the clustering result and combined with the line topology structure.

[0021] In order to achieve the above-mentioned purpose, the third aspect of the present invention provides a computer device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; the memory is used to store computer programs; the processor is used to implement the steps of the distribution network section positioning method based on multi-dimensional feature clustering of fault phase current as described in the first aspect when executing the program stored in the memory.

[0022] In order to achieve the above-mentioned objectives, the fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed by one or more processors, enables the processors to perform the steps of the distribution network section positioning method based on multi-dimensional feature clustering of fault phase current as described in the first aspect above.

[0023] In order to achieve the above-mentioned purpose, the fifth aspect of the present invention provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the distribution network section positioning method based on multi-dimensional feature clustering of fault phase current as described in the first aspect above.

[0024] The beneficial effects of the present invention are: The present invention uses the box counting dimension method to measure and compare the complexity of the three-phase transient current waveform, which can accurately identify the fault phase without setting a threshold. Based on the line selection results of the main station, the 7-dimensional data such as root mean square and peak factor are used to characterize the difference in transient current waveform of the fault phase upstream and downstream of the single-phase grounding fault point from multiple angles. The fuzzy C-means clustering is used to complete the classification of the fault section and the non-fault section, and the fault location is determined based on the clustering results. It is still effective under complex measurement conditions such as noise interference, delayed start of recording device, asynchronous sampling, random loss of sampling data, etc., and has strong anti-interference ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a flow chart of a method for locating a distribution network section based on multi-dimensional feature clustering of fault phase current according to the present invention; Figure 2 Schematic diagram of grid division of transient current curve and corresponding sub-matrix Figure 3 It is a topological structure diagram of a low-resistance grounding system in a verification embodiment; Figure 4 It is a schematic diagram of the fault feeder topology structure and fault location in the verification embodiment. DETAILED DESCRIPTION

[0026] The technical solution of the present invention is further described in detail below through specific implementation methods.

[0027] For ease of understanding, the interactive parties and / or terms and / or custom words involved in the present invention are first described in conjunction with the technical solution of the present invention: FCM, fuzzy C-means clustering, is an unsupervised "soft clustering" method that optimizes the objective function to obtain the membership of each sample point to all class centers, thereby determining the class of the sample point to achieve the purpose of automatically classifying the sample data. Each sample is assigned a membership function belonging to each cluster, and the samples are classified according to the size of the membership value.

[0028] The present invention proposes a method for locating the distribution network section based on multi-dimensional feature clustering of fault phase current. First, the transient current feature differences between the healthy phase and the fault phase of the fault feeder and the fault current flow path differences between the healthy section and the fault section are analyzed; then, based on the line selection result of the master station, the box counting dimension method is used to compare the waveform complexity differences of the three-phase transient current at the head end of the fault line to achieve fault phase identification; on this basis, multi-dimensional time domain feature extraction is performed on the fault phase transient current collected by each feeder terminal unit of the fault line, and fuzzy C-means clustering is performed on the feature data to complete the section positioning.

[0029] Example 1 This embodiment provides a distribution network section location method based on multi-dimensional feature clustering of fault phase current, such as Figure 1 As shown, the following steps are included: Step 1: When a single-phase grounding fault occurs, the three-phase current recording data at the head end of the fault line is collected according to the line selection result of the master station, and the current of one cycle after the fault is subtracted from the current of one cycle before the fault to eliminate the load current and construct the three-phase transient current.

[0030] Step 2: Use the box counting method to process the three-phase transient current, identify the fault phase, and upload the phase selection result to the master station.

[0031] Specifically, the method of processing the three-phase transient current by using the box counting dimension method to identify the fault phase includes the following steps: (1) The three-phase transient current waveform curve at the head end is binarized to obtain a binary image; the binary image can highlight the high-frequency oscillation characteristics of the transient current in a specific period of time and prevent the characteristics from being masked by amplitude differences.

[0032] It can be understood that before the binarization process is performed, the current value may be normalized first.

[0033] It should be noted that when the binary image is obtained, the binary matrix P corresponding to the binary image is also obtained, where element 0 represents the pixel point occupied by the waveform sampling point, and element 1 represents the blank area.

[0034] (2) Grid partitioning parameter setting: The grid size is dynamically adjusted in an exponential decreasing manner, and the matrix order corresponding to the binary image is adjusted to a power of 2.

[0035] Compared with the arithmetic decreasing method with a fixed step length, the exponential decreasing method can significantly shorten Approaching the minimum grid size in time and reducing the amount of calculation.

[0036] Furthermore, by adjusting the matrix order corresponding to the binary image to a power of 2, it can be ensured that each pixel of the binary image can be completely covered, thus preventing the edge of the curve from being missed.

[0037] In one embodiment, the matrix order N of the binary image is set to 2 to the 9th power, that is, 512×512, so as to take into account both the operation speed and the grid division accuracy.

[0038] (3) Mesh division: from the maximum mesh size Starting from, divide the binary matrix P into Sub-matrices, count the number of non-all-1 matrices ; then order ,statistics , and so on, until the minimum grid size is reached .

[0039] The third grid size is For example, the schematic diagram of the grid division result of the transient current curve and the corresponding sub-matrix are as follows Figure 2 shown. Figure 2 middle, NE It is a non-all-1 matrix, which means that there are pixels in this area of ​​the image, that is, there are 0 elements; E is a matrix of all 1s, indicating that the area does not contain any pixels. According to statistics, there are 10 non-all-1 matrices among the 16 sub-matrices, namely .

[0040] (4) Box counting dimension calculation: Finally, and The data series is fitted by the least squares straight line, and the slope is the box-counting dimension. D , where the phase transient current waveform box dimension is D The smallest one is the fault phase.

[0041] Step 3: The master station collects the fault phase current recording data of the detection points in each section of the fault line according to the phase selection result and extracts the seven-dimensional time domain features.

[0042] Specifically, after a high-resistance grounding fault occurs, the transient current waveforms of the fault phases in the fault section and the healthy section are greatly different in amplitude, sampling point distribution, waveform morphology, etc. In order to quantitatively describe the waveform difference characteristics, this embodiment refers to the commonly used time domain feature extraction method, and on the basis of the line selection results of the master station and the phase selection results of the box counting dimension, the 6-dimensional time domain features such as the root mean square of the transient current of the fault phases in each section of the fault line are constructed, and the dynamic time bending distance (DTW) between the fault phase current at each detection point and the fault phase current waveform at the head end of the line is calculated to form the 7th dimension feature, as shown below: Feature 1: Root Mean Square

[0043] Where: x i The current i sampling points, n is the total number of sampling points in the sampling period.

[0044] Feature 2: Peak-to-peak value

[0045] Feature 3: Crest Factor

[0046] Feature 4: Pulse Factor

[0047] Feature 5: Margin Factor

[0048] Feature 6: Number of zero crossings

[0049] in:

[0050] Feature 7: DTW between the fault phase current waveform at each detection point and the fault phase current waveform at the line head end The DTW algorithm is used to characterize the similarity between time series. Its advantage is that it supports multiple dynamic adjustments to the matching relationship between data to find the maximum similarity between two sequences. The higher the similarity between the sequences, the smaller the dynamic time warping distance and the better the ability to resist synchronization errors.

[0051] For the fault phase current sequence at the head end of the fault lineP ={ p 1, p 2,…… p n} and the fault phase current sequence of the healthy section Q ={ q 1, q 2,…… q n}, first construct the cumulative distance matrix O ,in:

[0052]

[0053] Then the cumulative distance matrix O Filling is performed, where:

[0054] Finally, the two phase current sequences P and Q The DTW distance is O ( n , n ).

[0055] In addition, since the dimensions of each dimensional data are not uniform, it is necessary to normalize the dimensional data and limit them to the range of 0~1, as shown in the following formula:

[0056] In the formula, T i Indicates i dimensional features, in this embodiment i =1…7; sT i Represents the normalized i Dimensional features, T i ( k ) indicates the k The phase current corresponding to the first detection point i Dimensional features.

[0057] Step 4: Perform FCM clustering based on the seven-dimensional time domain features, and determine the fault location based on the clustering results and the line topology.

[0058] Specifically, FCM clustering based on seven-dimensional time domain features includes the following steps: Setting the objective function J for:

[0059]

[0060] Among them, 2 is the number of cluster centers, including two cluster centers of the fault section and the sound section; γ ik is the membership degree, which is used to characterize the i The probability that a detection point belongs to a faulty section or a sound section; n is the total number of current detection points on the fault line; m is the weight parameter used to control γ ik The degree of fuzzification, in this embodiment, m is preferably 2; y i Indicates i The transient current characteristic vector of the fault phase at each detection point; μ k Indicates k The feature vector corresponding to the cluster center has the same dimension as x i consistent;|| y i - μ k || represents the Euclidean distance, which is used to measure the feature vector y i and μ k The smaller the Euclidean distance, the closer the i The closer the detection point is to the k Cluster centers; Use the FCM algorithm to find the objective function through multiple iterations J The minimum value of μ k and membership γ ik To perform synchronous update, the formula is as follows:

[0061]

[0062] After updating, continue to calculate the objective function value until convergence to obtain the clustering result.

[0063] Since the fault current will definitely flow through the head end of the faulty line when a single-phase grounding fault occurs, the category to which the detection point at the head end of the faulty line belongs can be determined as the fault section, and the other category is the sound section.

[0064] It can be understood that after determining the location of the fault, the master station issues a fault isolation command to the distribution network, disconnects the circuit breakers at both ends of the section where the fault point is located, and completes the fault isolation work.

[0065] Verification example: In order to verify the effectiveness of the method proposed in this embodiment, a simulation model of a 10kV distribution network low-resistance grounding system is built based on the Matlab / Simulink simulation platform. Figure 3 shown.

[0066] In the simulation model, the upstream 110kV power supply line is replaced by a 110kV AC power supply; the main transformer is set to a 110kV / 10.5kV Y-△ type transformer; to lead out the neutral point, a Z-type grounding transformer is connected to the secondary side of the main transformer, and a neutral point grounding resistor with a resistance of 10Ω is connected; the 10kV bus has 7 lines, of which lines 1, 2, and 4 are pure cable lines, lines 5 and 6 are pure overhead lines, lines 3 and 7 are cable mixed lines, and line 7 is set to contain two branch lines, and its line impedance parameters are shown in Table 1: Table 1 Line impedance parameters

[0067] Since the implementation time point of the method in this embodiment is after the fault line identification process, for the control variable, it is assumed that the fault uniformly occurs in the cable hybrid line L7, where the detailed line topology structure of L7 is as follows: Figure 3 shown.

[0068] In the figure, the dotted ellipse represents the phase current detection point on the feeder, and F1~F4 represent the fault location, which are used to simulate the feeder trunk, feeder end, branch line fault, etc. The fault point F1 is 6km away from the line head; the fault point F2 is at the end of the line, 20km away from the line head; the fault point F3 is located at the branch line II, 6km away from the branch point; the fault point F4 is at the end of the branch line I, 12km away from the branch point.

[0069] Different fault conditions are set at F1~F4, where different transition resistances are set at fault point F1, different fault phases are set at fault point F2, and different fault phase angles are set at fault point F3. At fault point F4, the ability of the positioning method to withstand transition resistance under a small fault phase angle at the end of the line is tested, and the sampling frequency is 10kHz. The simulation results are shown in Table 2: Table 2 Fault phase identification and section location results under different fault conditions , It can be seen from the simulation results that the method proposed in this embodiment can reliably identify the fault phase and accurately locate the section where the grounding fault point is located under different fault positions, different transition resistances, different fault phase angles, and different fault phases; and is suitable for complex fault conditions such as line end faults, small fault phase angle faults, and branch line faults. In the most unfavorable case of a small phase angle fault at the end of a branch line, the method proposed in this embodiment can withstand a transition resistance of up to 7000Ω, with high sensitivity.

[0070] 1) Study on the adaptability of the method under noise signal interference In order to verify the positioning accuracy of the method in this embodiment under the interference of noise signals, a 3kΩ high-resistance grounding fault of phase B occurs at F4 of branch line I of feeder 7, and the fault phase angle is 0°. In the current sampling data of the detection points of each section, a Gaussian white noise signal of 10-40dB is added to the normal operation period and the grounding fault period, respectively. The phase selection and positioning results are shown in Table 3: Table 3 Results of fault phase identification and section location under different signal-to-noise ratio conditions , It can be seen from Table 3 that with the decrease of the signal-to-noise ratio, a large number of high-frequency components are mixed into the sampled signal, resulting in an increase in the complexity of the three-phase transient current image. The corresponding three-phase box-counting dimensions are increased, and the box-counting dimension distinction between the fault phase and the non-fault phase is also reduced. However, the method proposed in this embodiment still has good noise adaptability, and can effectively identify the fault phase and fault section under the condition of 10-40dB signal-to-noise ratio.

[0071] 2) Analysis of method adaptability under delayed sampling and asynchronous sampling When a single-phase grounding fault occurs, the time when the fault recording equipment at different locations switches from the sleep state to the receiving state may be different. In order to verify the effectiveness of the method proposed in this embodiment under the condition of delayed sampling and asynchronous sampling, a B-phase 3kΩ high-resistance grounding fault is set at F4 of branch line I of feeder 7, and the fault phase angle is 0°; the power industry standard stipulates that the trigger delay of the recording equipment is not higher than 0.5ms and the three-phase synchronization error is not higher than 0.1ms. This section sets the delay time tA, tB, and tC of the three-phase current acquisition time relative to the fault time to be much larger than 0.1ms to verify the effectiveness of the phase selection method.

[0072] The simulation results are shown in Table 4: Table 4 Phase selection results under delayed sampling and asynchronous sampling , On this basis, the fault phase current collected by some detection points is randomly set to be delayed at different times at the fault moment to simulate the situation where the recording equipment at the detection points in each section is started late and the sampling is not synchronized. The set delay time is much larger than the delay time specified in the standard (0.5ms).

[0073] The simulation results are shown in Table 5: Table 5 Segment positioning results under delayed sampling and asynchronous sampling , It can be seen from the simulation results that no matter whether it is the synchronous delayed sampling of the three-phase current or the asynchronous delayed sampling of the three-phase current, the method proposed in this embodiment can accurately identify the fault phase; when there are different degrees of sampling delays at detection points in different sections, the method proposed in this embodiment can still accurately locate the fault position and has strong adaptability.

[0074] 3) Analysis of method adaptability under random missing sampling point data There may be a problem of random sampling point loss in the process of the distribution line measurement equipment transmitting data to the main grid. To simulate this phenomenon, a B-phase 3kΩ high-resistance grounding fault is set at F4 of the branch line I of the feeder 7, and the fault phase angle is 0°. A certain proportion of the current waveform sampling points are randomly set to zero, and the effectiveness of the method proposed in this embodiment is verified in this case. The simulation results are as follows: Table 6 Phase selection results when a certain proportion of data is randomly missing , Table 7 Segment location results when data is randomly missing at a certain ratio ,

[0075] It can be seen from the simulation results that the method in this embodiment is still accurate and has good robustness when the sampling data is lost.

[0076] 4.6 Analysis of method adaptability at different sampling frequencies In the case of a single-phase grounding fault, the natural frequency of the fault transient current is generally between 0.3 and 3kHz. The State Grid Corporation standard stipulates that the sampling frequency of the transient recording device shall not be less than 4kHz. According to Shannon's theorem, at a sampling frequency lower than 6kHz, the waveform data will be distorted. In order to verify the effectiveness of the method in this embodiment at different sampling frequencies, a C-phase 3kΩ high-resistance grounding fault is set at F3 at the II of the feeder 7 branch line, and the fault phase angle is 60°. The simulation results at different sampling frequencies are as follows: Table 8 Phase selection results at different sampling frequencies

[0077] Table 9 Segment positioning results at different sampling frequencies , It can be seen from the simulation results that at a low sampling frequency of 1kHz, there are only 20 sampling points in a single cycle, and the transient currents of the three phases and each section are seriously distorted. The waveform characteristics are no longer different, the difference in the three-phase box dimensions is gradually reduced, and mislocation occurs. However, for sampling frequencies of 4kHz and above in actual engineering, the method in this embodiment can accurately identify the faulty phase and faulty section.

[0078] It should be understood that, although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in the present embodiment, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the above embodiments may include a plurality of steps or a plurality of stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of the steps or stages in other steps.

[0079] Example 2 Based on the same inventive concept, the embodiment of the present application also provides a distribution network section positioning device based on multi-dimensional feature clustering of fault phase current for implementing the distribution network section positioning method based on multi-dimensional feature clustering of fault phase current involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the distribution network section positioning device based on multi-dimensional feature clustering of fault phase current provided below can refer to the limitations of the distribution network section positioning method based on multi-dimensional feature clustering of fault phase current above, and will not be repeated here.

[0080] Specifically, Figure 4 As shown, the distribution network section positioning device based on multi-dimensional feature clustering of fault phase current includes: The three-phase current acquisition module is used to collect the three-phase current recording data at the head end of the fault line according to the line selection result of the master station when a single-phase grounding fault occurs, and the current of one cycle after the fault is subtracted from the current of one cycle before the fault to eliminate the load current and construct the three-phase transient current; The fault phase screening module is used to process the three-phase transient current using the box counting dimension method, identify the fault phase, and upload the phase selection result to the main station; The feature extraction module is used by the master station to collect the fault phase current recording data of the detection points in each section of the fault line according to the phase selection result, and extract the seven-dimensional time domain features; Clustering module, used for FCM clustering based on seven-dimensional time domain features; The fault determination module is used to determine the fault location based on the clustering result and combined with the line topology structure.

[0081] Example 3 The present application embodiment provides a computer device, which may be a terminal. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. The processor, the memory, and the input / output interface are connected via a system bus, and the communication interface, the display unit, and the input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, an XX method is implemented. The display unit of the computer device is used to form a visually visible picture, which may be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0082] Example 4 On the basis of the above embodiments, this embodiment provides a readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented.

[0083] Example 5 Based on the above embodiments, this embodiment provides a computer program product, including a computer program. When the computer program is executed by a processor, the following steps are implemented.

[0084] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0085] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0086] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or some technical features can be replaced by equivalents without departing from the spirit of the technical solution of the present invention, which should be included in the scope of the technical solution for protection of the present invention.

Claims

1. A distribution network section location method based on multi-dimensional feature clustering of fault phase current, characterized in that: The following steps are involved: When a single-phase grounding fault occurs, the three-phase current recording data at the head end of the fault line is collected according to the line selection result of the master station, and the current of one cycle after the fault is subtracted from the current of one cycle before the fault to eliminate the load current and construct the three-phase transient current; The three-phase transient current is processed using the box counting dimension method to identify the fault phase, and the phase selection result is uploaded to the main station; Based on the phase selection results, the master station collects the fault phase current recording data of the detection points in each section of the fault line and extracts the seven-dimensional time domain features; FCM clustering is performed based on seven-dimensional time domain features, and the fault location is determined based on the clustering results and combined with the line topology.

2. A distribution network section location method based on multi-dimensional feature clustering of fault phase current according to claim 1, characterized in that: The method of processing the three-phase transient current by using the box counting dimension method to identify the fault phase includes the following steps: Binarization processing is performed on the three-phase transient current waveform curve at the head end; The grid size is dynamically adjusted in an exponential decreasing manner, and the matrix order corresponding to the binary image is adjusted to a power of 2; From the maximum grid size Starting from, divide the binary matrix P into Sub-matrices, count the number of non-all-1 matrices ; then order ,statistics , and so on, until the minimum grid size is reached ; Last pair and The data series is fitted by the least squares straight line, and the slope is the box-counting dimension. D , where the phase transient current waveform box dimension is D The smallest one is the fault phase.

3. A distribution network section location method based on multi-dimensional feature clustering of fault phase current according to claim 1 or 2, characterized in that: The seven-dimensional time domain features include: the dynamic time bending distance between the fault phase current at each detection point and the fault phase current waveform at the head end of the line, and the root mean square, peak-to-peak value, peak factor, pulse factor, margin factor, and number of zero crossings of the transient current of the fault phase in each section of the fault line.

4. A distribution network section location method based on multi-dimensional feature clustering of fault phase current according to claim 3, characterized in that: FCM clustering is performed based on seven-dimensional time domain features, including: Setting the objective function J for: ， ， Among them, 2 is the number of cluster centers, including two cluster centers of the fault section and the sound section; γ ik is the membership degree, which is used to characterize the i The probability that a detection point belongs to a faulty section or a sound section; n is the total number of current detection points on the fault line; m is the weight parameter used to control γ ik The degree of blurring; y i Indicates i The transient current characteristic vector of the fault phase at each detection point; μ k Indicates k The feature vector corresponding to the cluster center has the same dimension as x i consistent;|| y i - μ k || represents the Euclidean distance, which is used to measure the feature vector y i and μ k The smaller the Euclidean distance, the closer the i The closer the detection point is to the k Cluster centers; Use the FCM algorithm to find the objective function through multiple iterations J The minimum value of μ k and membership γ ik To perform synchronous update, the formula is as follows: ， ， After updating, continue to calculate the objective function value until convergence to obtain the clustering result.

5. A distribution network section location method based on multi-dimensional feature clustering of fault phase current according to claim 4, characterized in that: The steps to obtain the dynamic time warping distance are: Set the fault phase current sequence at the head end of the fault line P ={ p 1, p 2,…… p n }, fault phase current sequence of healthy section Q ={ q 1, q 2,…… q n }; Constructing cumulative distance matrix O ,in: ， ， Cumulative distance matrix O Filling is performed, where: ， Finally, the two phase current sequences P and Q The DTW distance is O ( n , n ).

6. A distribution network section positioning device based on multi-dimensional feature clustering of fault phase current, characterized in that: include: The three-phase current acquisition module is used to collect the three-phase current recording data at the head end of the fault line according to the line selection result of the master station when a single-phase grounding fault occurs, and the current of one cycle after the fault is subtracted from the current of one cycle before the fault to eliminate the load current and construct the three-phase transient current; The fault phase screening module is used to process the three-phase transient current using the box counting dimension method, identify the fault phase, and upload the phase selection result to the main station; The feature extraction module is used by the master station to collect the fault phase current recording data of the detection points in each section of the fault line according to the phase selection result, and extract the seven-dimensional time domain features; Clustering module, used for FCM clustering based on seven-dimensional time domain features; The fault determination module is used to determine the fault location based on the clustering result and combined with the line topology structure.

7. A computer device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; The processor is used to implement the distribution network section positioning method based on multi-dimensional feature clustering of fault phase current as described in any one of claims 1 to 5 when executing the program stored in the memory.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the distribution network section locating method based on multi-dimensional feature clustering of fault phase current described in any one of claims 1 to 5 are implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the distribution network section locating method based on multi-dimensional feature clustering of fault phase current described in any one of claims 1 to 5 are implemented.

Images