Power distribution network single-phase earth fault single-end traveling wave fault location method
By combining the improved Pettitt test, Pearson correlation coefficient and Hough transform, the error problem of single-phase grounding fault location in complex distribution networks is solved, and high-precision and low-cost fault location is achieved.
Patent Information
- Application Number
- CN202510124695.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-26
- Publication Date
- 2025-09-16
AI Technical Summary
The existing single-phase grounding fault location method in distribution networks is easily affected by the interference of reflected wave heads in complex networks, resulting in large ranging errors. In addition, the data storage and processing pressure brought by the high sampling rate affects the positioning efficiency, making it difficult to accurately locate the fault point in complex topology structures.
The improved Pettitt test method is used to analyze the statistical characteristics of waveform data. The Pearson correlation coefficient and Hough transform are combined to identify the arrival time of the fault wave head. The wave head position is accurately calibrated by identifying the intersection of straight lines in the waveform, reducing the interference of reflected wave heads and improving positioning accuracy and noise resistance.
Accurately calibrate the reflected wave head at the fault point in complex distribution networks, reduce ranging errors, improve positioning accuracy and anti-interference capabilities, reduce hardware costs, and achieve efficient and reliable fault location.
Smart Images

Figure CN120652203A_ABST
Abstract
Claims
1. A single-ended traveling wave distance measurement method for single-phase ground fault in a distribution network, characterized by: The method accurately identifies the arrival time of the ground fault traveling wave front by calculating the test statistic of the traveling wave waveform data, avoiding the ranging error caused by inaccurate fault time, thereby improving the fault location accuracy. At the same time, by analyzing the correlation between the original waveform and the reference waveform, the positions of the first wave front and the reflected wave front are accurately identified, reducing the influence of the reflected wave front on the positioning accuracy. In addition, by extracting the straight line features in the waveform and using the intersection of the two straight lines to accurately calibrate the arrival time of the wave front.
2. A single-ended traveling wave distance measurement method for single-phase ground fault in distribution network according to claim 1, characterized in that: The method comprises the following steps: Step S1: After a single-phase ground fault occurs, the time window where the fault wave head is located is accurately determined by analyzing the test statistics of the distribution line waveform data, thereby avoiding inaccurate ranging caused by incorrect calibration of the wave head arrival time and improving the accuracy of wave head identification; Step S2: After determining the time window where the fault wave head is located, analyze the similarity between the original waveform and the reference waveform to accurately locate the time window of the traveling wave head and the reflected wave head at the fault point to reduce the interference of other reflected wave heads on positioning and improve the reliability of ranging; Step S3: After determining the locations of the first wave head and the reflected wave head, identify the straight lines contained in the traveling wave waveform, and calibrate the arrival time of the wave head by identifying the intersection of the straight lines, thereby improving the accuracy of fault location and achieving more accurate ranging results.
3. The single-ended traveling wave distance measurement method for single-phase ground fault in distribution network according to claim 2, characterized in that: In step S1, the fault moment is captured, and the mutation point in the detection waveform is determined by calculating the rank difference between the front and back parts of the waveform and maximizing the difference. The method traverses the waveform data in the time window, calculates the test statistic for each data point, and compares the rank difference between the front and back subsequences. After finding the maximum statistic, the significance level value is calculated, which represents the probability of observing the maximum statistic under the assumption that there is no mutation point. If it is lower than a specific threshold, it is considered that a significant mutation point exists.
4. The single-ended traveling wave distance measurement method for single-phase ground fault in distribution network according to claim 2, characterized in that: In step S2, the degree of linear correlation between the two waveforms is measured to determine the position where the original waveform is most similar to the reference waveform; The specific steps are: traverse the waveform data within the time window, select historical fault data as the reference waveform, and calculate the correlation coefficient around each data point; Match this reference waveform with the original waveform. When the correlation coefficient reaches the maximum value, it indicates that the two waveforms are most similar. The matching position at this time is the target wave head position.
5. The single-ended traveling wave distance measurement method for single-phase ground fault in distribution network according to claim 2, characterized in that: In step S3, the arrival time of the wave head is calibrated; this method uses feature extraction technology and point-line duality to map the sampling points in the image space to the intersection of the curves in the parameter space. The intersection of the curves represents the parameters of the line in the image; in order to adapt to one-dimensional waveform detection: first, the parameter space is discretized and divided into grids to construct the Hough matrix; then, the cumulative number of sampling points in each grid of the parameter space is counted, and the maximum value point in the cumulative array is found, which corresponds to the possible line parameters; then, other maximum value points are found to represent other possible lines.