A method for locating traveling waves in power distribution lines based on multi-factor analysis
By combining parallel computing and error propagation law evaluation with multi-factor analysis, the traveling wave localization method for power distribution lines solves the problem of decoupling error factors in existing technologies, achieving high-precision fault location and efficiency improvement.
Patent Information
- Application Number
- CN202611123337.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-28
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2046-07-28
AI Technical Summary
Existing power distribution line traveling wave positioning technology cannot decouple four types of error factors: line length, traveling wave speed, clock synchronization, and wavehead identification time. It is difficult to quantify the influence ratio of each error source, resulting in poor positioning results, lack of effective verification basis, insufficient positioning accuracy, and low fault handling efficiency.
Parallel computation is employed using the dual-end traveling wave time difference method, the multi-end traveling wave positioning method, and the single-end traveling wave positioning method. Combined with the error propagation law and data quality assessment, the fault location range is generated and the dominant error source is identified through weighted fusion of algorithm consistency score, comprehensive data quality score, and error control confidence.
It enables precise tracing of error sources and real-time assessment of data quality, improves positioning accuracy and fault handling efficiency, and provides a reliability and confidence evaluation system for positioning results.
Smart Images

Figure CN122632011B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power distribution line location technology, and in particular to a method for locating traveling waves in power distribution lines based on multi-factor analysis. Background Technology
[0002] Traveling wave localization of distribution lines is a mainstream technique for power grid fault diagnosis. Currently, most conventional traveling wave localization schemes use a single localization algorithm to determine the fault location, without conducting a comprehensive quality assessment of the traveling wave data collected on-site, or cross-validating the results of multiple algorithms. Under this single-algorithm operation mode, the localization results lack effective verification criteria, resulting in poor stability of the output.
[0003] Existing technologies generally lack systematic decomposition and quantitative analysis of positioning errors. Multiple error factors, such as line length, traveling wave velocity, measurement terminal clock synchronization, and wave front arrival time, will jointly affect the positioning results. Existing methods cannot calculate the contribution ratio of various errors, make it difficult to determine the reliability of the results, and cannot generate reasonable fault location intervals, resulting in low overall fault location accuracy and difficulty in meeting the actual operation and maintenance requirements of power distribution lines. Summary of the Invention
[0004] In view of the above problems, the purpose of this invention is to provide a traveling wave location method for power distribution lines based on multi-factor analysis, so as to solve the problems that traditional traveling wave ranging technology for power distribution lines cannot decouple four types of error factors: line length, traveling wave velocity, clock synchronization, and wavefront identification time, making it difficult to quantify the influence ratio of each error source and achieve error tracing. At the same time, it solves the problems that existing technologies lack real-time data quality assessment, multi-dimensional cross-validation, and a confidence evaluation system for location results. When the signal is interfered with, it is easy to produce biased results, and it cannot rely on redundant information to complete error compensation and targeted parameter calibration, which leads to a lack of direction, insufficient location accuracy, and low efficiency in on-site fault diagnosis.
[0005] This invention provides a method for locating traveling waves in power distribution lines based on multi-factor analysis, comprising: Based on the fault traveling wave data collected from the power distribution line after quality assessment, parallel calculations were performed using the double-ended traveling wave time difference method, the multi-ended traveling wave location method, and the single-ended traveling wave location method to obtain the preliminary location results of the power distribution line. Statistical analysis is performed on the standard deviation or coefficient of variation of the dispersion between the preliminary positioning results to obtain the algorithm consistency score of the effective positioning results; Based on the error propagation law, the independent positioning error components of the current power distribution line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error are evaluated respectively, and the error contribution of each error source and the overall positioning error are obtained. The algorithm consistency score, the comprehensive data quality score of the fault traveling wave data, and the error control confidence obtained from the comprehensive positioning error mapping are weighted and fused to obtain the comprehensive confidence of the effective positioning result. The confidence level is determined based on the overall confidence level, and the location range of the fault point in the power distribution line is generated based on the overall location error. At the same time, the dominant error source with the largest contribution among all error sources is output.
[0006] Preferably, the process of using fault traveling wave data collected from power distribution lines and subjected to quality assessment is as follows: The wavefronts of the three-phase current traveling wave waveforms synchronously acquired by multiple measurement terminals of the power distribution line are detected by wavelet transform modulus maxima method. The confidence level of wavefront identification is calculated based on the ratio of the wavefront energy of the traveling wavefront to the total energy within the preset analysis window. At the same time, the signal-to-noise ratio of the same waveform is calculated and mapped to the signal quality score. The wavefront identification confidence score and the signal quality score are weighted and fused to obtain a comprehensive data quality score; Traveling wave data with a comprehensive data quality score higher than a preset quality threshold are identified as faulty traveling wave data that have undergone quality assessment.
[0007] Preferably, the preliminary location results of the power distribution line are obtained by parallel calculation using the two-end traveling wave time difference method, the multi-end traveling wave positioning method, and the single-end traveling wave positioning method. The process is as follows: Based on the waveform of the same fault event in the fault traveling wave data that has been quality assessed, three localization algorithms are activated simultaneously; among them, the double-end traveling wave time difference method uses the difference in arrival time of the wavefront of the measurement terminals at both ends of the line and the line length and traveling wave speed to calculate the first fault distance; The multi-terminal traveling wave positioning method uses the combination of wavefront arrival times of three or more measurement terminals to establish a system of simultaneous equations to solve for the second fault distance; The single-ended traveling wave location method calculates the third fault distance using the time delay difference between the reflected wave at the fault point and the initial traveling wave surge. The first fault distance, the second fault distance, and the third fault distance are used together as the preliminary location result.
[0008] Preferably, the process of statistically analyzing the standard deviation or coefficient of variation of the dispersion among the preliminary positioning results to obtain the algorithm consistency score of the effective positioning results is as follows: Based on at least two valid localization results obtained from parallel computing, calculate the arithmetic mean of all valid localization results; The standard deviation of the valid positioning results is generated based on the arithmetic mean and each valid positioning result; Multiply the ratio of the standard deviation to the arithmetic mean by 100% to obtain the coefficient of variation of the effective localization results; The coefficient of variation is compared with a preset consistency threshold, and an algorithm consistency score is generated for the effective positioning result according to the rule that the smaller the coefficient of variation, the higher the consistency score.
[0009] Preferably, the formula for calculating the consistency score of the algorithm is as follows: ; In the formula, The algorithm consistency score is given, with a value range of [0,1]. It is a function with maximum value. The coefficient of variation is 1. This is the consistency threshold.
[0010] Preferably, the independent positioning error components of the current power distribution line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error are evaluated according to the error propagation law to obtain the error contribution of each error source and the overall positioning error. The process is as follows: The estimated values of the current power distribution line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error are obtained. The estimated values are obtained through iterative statistics based on historical fault data. According to the law of error propagation, calculate the positioning error component caused by each error source acting alone: multiply the estimated line length error by one-half to obtain the length error positioning component; Multiply the estimated traveling wave velocity error by half the time difference in the two-end traveling wave time difference method to obtain the wave velocity error positioning component; The synchronization error positioning component is obtained by multiplying the estimated value of the measurement terminal synchronization clock error by half of the traveling wave velocity; the wavefront arrival time error positioning component is obtained by multiplying the estimated value of the wavefront arrival time error by half of the traveling wave velocity. The overall positioning error is calculated using an error synthesis formula, which is: ; In the formula, To account for the overall positioning error, To locate the component of length error, To locate the wave velocity error component, To locate the synchronization error component, The wavefront timing error is located as a component.
[0011] Preferably, the process of obtaining the error contribution of each error source is as follows: Calculate the square values of the length error positioning component, wave velocity error positioning component, synchronization error positioning component, and wavefront timing error positioning component respectively; Summing up all the squared values yields the sum of squares. Then, each squared value is calculated as a percentage of the sum of squares. Each percentage is multiplied by 100% to convert it into a corresponding percentage, thus obtaining the contribution of line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error. Each error contribution percentage is used to quantitatively identify the dominant error source of the current positioning error, and the error source with the highest contribution percentage is determined as the dominant error source.
[0012] Preferably, the error control confidence level obtained from the comprehensive positioning error mapping is achieved through the following process: The overall positioning error is compared with the preset error control threshold in real time; When the overall positioning error does not exceed the error control threshold, the error control confidence level is calculated according to the ratio of the overall positioning error to the error control threshold, and the error control confidence level decreases monotonically as the overall positioning error increases. When the overall positioning error exceeds the error control threshold, the error control confidence level is set to the minimum value; the range of the error control confidence level is limited to a preset value range.
[0013] Preferably, the process of weightedly fusing the algorithm consistency score, the comprehensive data quality score of the fault traveling wave data, and the error control confidence score obtained from the comprehensive positioning error mapping to obtain the comprehensive confidence score of the effective positioning result is as follows: Each of the following is assigned a preset weight coefficient: algorithm consistency score, comprehensive data quality score, and error control confidence score. The sum of the weight coefficients is one. The weighted fusion result is obtained by adding the product of the algorithm consistency score and its corresponding weight coefficient, the product of the comprehensive data quality score and its corresponding weight coefficient, and the product of the error control confidence score and its corresponding weight coefficient, and is used as the comprehensive confidence score. The range of the overall confidence level is limited to a preset numerical range, and the higher the value, the higher the reliability of the preliminary positioning result.
[0014] Preferably, the process of determining the confidence level based on the comprehensive confidence level, generating the location range of the fault point in the power distribution line based on the comprehensive location error, and simultaneously outputting the identifier of the dominant error source with the largest contribution among all error sources is as follows: The real-time overall confidence level is compared with multiple preset confidence threshold ranges to determine the confidence level to which the overall confidence level belongs. The confidence levels, from high to low, include at least extremely high confidence, high confidence, medium confidence, low confidence, and insufficient confidence. Using the arithmetic mean of the preliminary positioning results as the positioning reference point, and the comprehensive positioning error as the unilateral offset, the distance of the comprehensive positioning error is offset to the left and right on both sides of the positioning reference point to generate the positioning range of the fault point in the power distribution line. Compare the error contribution of each error source, identify the error source with the largest error contribution as the dominant error source, and output the identification information of the dominant error source.
[0015] As can be seen from the above technical solution, the present invention provides a method for locating traveling waves in power distribution lines based on multi-factor analysis. The present invention can quantify the influence weights of four types of factors on the positioning error: line length, traveling wave velocity, clock synchronization, and wavefront identification time, so as to achieve accurate error tracing. At the same time, it can evaluate data quality in real time, analyze algorithm consistency and classify positioning information level, clarify the reliability of the results, and complete parameter correction and error compensation based on the error contribution, effectively reducing the comprehensive positioning error, significantly improving the fault location accuracy of power distribution lines, and promoting the overall improvement of the efficiency and intelligence level of fault handling work. Attached Figure Description
[0016] Other objects and results of the invention will become more apparent and readily understood by referring to the following description taken in conjunction with the accompanying drawings, and with a more complete understanding of the invention. In the drawings: Figure 1 This is a schematic diagram illustrating the principles of reflection and refraction during the propagation of traveling waves. Figure 2 A schematic diagram illustrating the principle of wave-shaped reflection positioning; Figure 3 The diagram below illustrates the actual cable line parameters and distribution network topology of an exemplary embodiment of the present invention. Figure 4 Simulation diagrams of arrival times and wave velocities for three traveling waves; Figure 5 This is a flowchart illustrating a method for locating traveling waves in power distribution lines based on multi-factor analysis, according to an embodiment of the present invention. Figure 6 This is a data graph of the traveling wave head detection results according to an embodiment of the present invention; Figure 7 This is a bar chart showing the data quality scoring results of an embodiment of the present invention; Figure 8 This is a schematic diagram of traveling wave velocity measurement according to an embodiment of the present invention; Figure 9 This is a diagram showing the consistency results of multi-algorithm localization in an embodiment of the present invention; Figure 10 This is a graph showing the contribution of error sources in an embodiment of the present invention. Figure 11 This is a graph showing the overall confidence level fusion result of an embodiment of the present invention; Figure 12 This is a schematic diagram of the fault location interval results in an embodiment of the present invention. Detailed Implementation
[0017] This invention primarily addresses the problem that traditional traveling wave ranging technology for power distribution lines cannot decouple four types of error factors: line length, traveling wave velocity, clock synchronization, and wavefront identification time. It also makes it difficult to quantify the impact ratio of each error source and achieve error tracing. Furthermore, it solves the problems of existing technologies lacking real-time data quality assessment, multi-dimensional cross-validation, and a confidence evaluation system for positioning results. These technologies are also prone to producing biased results when the signal is interfered with, and cannot rely on redundant information to complete error compensation and targeted parameter calibration. Consequently, these technologies result in a lack of direction in on-site fault diagnosis, insufficient positioning accuracy, and low efficiency in fault handling.
[0018] To address the aforementioned problems, this invention provides a method for locating traveling waves in power distribution lines based on multi-factor analysis. The specific embodiments of this invention will be described in detail below with reference to the accompanying drawings.
[0019] A cable line is a uniformly distributed parameter element. When a traveling wave propagates along the line, its impedance remains constant. However, when the traveling wave reaches the connection point of other electrical equipment, the circuit parameters change abruptly, and the impedance changes accordingly. The linear relationship established between the traveling wave voltage and current on the line is disrupted. At this point, part of the traveling wave returns to the cable line, while another part is transmitted through the connection point into other circuits. This phenomenon is called reflection and refraction of traveling waves. These traveling waves obey Kirchhoff's laws at the connection point, from which the reflection coefficient and refraction coefficient at the connection point can be calculated. Figure 1 For example, A traveling wave propagates from one line to the other. The wave impedance of the incident line is Z1, and the wave impedance of the refracted line is Z2. Assuming the direction of the incident wave is positive, the incident wave is a positive traveling wave, the reflected wave is a negative traveling wave, and the refracted wave is a positive traveling wave. The principle of traveling wave reflection and refraction positioning is as follows: Figure 2 As shown, by sequentially performing pulse calibration, feature extraction, pulse classification, and reflection point property identification on the acquired traveling wave signal, the traveling wave refraction and reflection distance can be calculated, thereby determining the location of the fault point.
[0020] To illustrate the traveling wave localization method for power distribution lines based on multi-factor analysis provided by this invention, the embodiments of this invention are as follows: Figure 3 This explanation uses a specific cable line parameter and distribution network topology as an example. Figure 3In the topology, the fault occurs outside the detection points. The traveling waves generated by the fault pass through detection points A, B, and C. By detecting the initial arrival times of the traveling waves at detection points A, B, and C, the wave velocity of the conductor can be calculated. Detection points A and C are on the main line, while detection point B is on a branch line. The faults at the line terminals B, C, and D are considered to be outside the designated area. The traveling waves at these three measurement points are captured, and the arrival times and velocities of the three traveling waves are calculated as follows: Figure 4 As shown.
[0021] Figure 5 An exemplary illustration is provided for a method for locating traveling waves in power distribution lines based on multi-factor analysis, according to an embodiment of the present invention.
[0022] The following description of exemplary embodiments is merely illustrative and is in no way intended to limit the invention or its application or use. Techniques and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques and equipment should be considered part of the specification.
[0023] Reference Figure 5 The diagram shown is a flowchart illustrating a method for locating traveling waves in distribution lines based on multi-factor analysis, according to an embodiment of the present invention. In this embodiment, the method for locating traveling waves in distribution lines based on multi-factor analysis includes: Step 1: Based on the fault traveling wave data collected from the power distribution line after quality assessment, parallel calculations are performed using the double-ended traveling wave time difference method, the multi-ended traveling wave location method, and the single-ended traveling wave location method to obtain the preliminary location results of the power distribution line. In this embodiment of the invention, the process of using fault traveling wave data collected from power distribution lines and subjected to quality assessment is as follows: The wavefronts of the three-phase current traveling wave waveforms synchronously acquired by multiple measurement terminals of the power distribution line are detected by wavelet transform modulus maxima method. Reliable fault traveling wave timing characteristics are obtained through multi-terminal synchronous acquisition and wavefront identification, such as... Figure 6 As shown. Figure 6 This diagram displays the fault traveling wave waveform and wavefront detection results synchronously acquired by multiple measurement terminals on a power distribution line. The horizontal axis represents time, and the vertical axis represents the normalized current amplitude. Different colored curves correspond to the traveling wave signals acquired by the measurement terminals at terminals M, T2, T3, and N, respectively. The red vertical line and red markers in the diagram indicate the arrival time of the traveling wave front identified using the wavelet transform modulus maxima method.
[0024] from Figure 6As can be seen, the arrival time of the transient traveling wave at different terminals varies significantly. Terminal T3 detects the wavefront earliest, indicating it is closest to the fault point; terminal M detects the wavefront arrival time latest, indicating it is relatively farther from the fault point. This result is consistent with the physical process of traveling waves propagating along both sides of a distribution line. The wavefront arrival time serves as the fundamental input for the two-terminal traveling wave time difference method, the multi-terminal traveling wave location method, and the single-terminal reflected wave location method, directly affecting the accuracy of subsequent fault distance calculations.
[0025] The confidence level of wavefront identification is calculated based on the ratio of the wavefront energy of the traveling wavefront to the total energy within the preset analysis window. At the same time, the signal-to-noise ratio of the same waveform is calculated and mapped to the signal quality score. Figure 7 The graph displays the wavefront identification confidence score, signal quality score, and overall data quality score for each measurement terminal. The red horizontal line represents a preset data quality threshold used to determine whether the acquired fault traveling wave data meets the location calculation requirements. The wavefront identification confidence score is calculated as the ratio of the wavefront signal energy to the total energy within the analysis window, reflecting the reliability of the wavefront detection results. The signal quality score is obtained by mapping the signal-to-noise ratio, reflecting the degree of noise interference on the acquired waveform. The overall data quality score is obtained through weighted fusion, for example, using a wavefront identification confidence score weight of 0.6 and a signal quality score weight of 0.4. Figure 7 As can be seen, the overall quality scores of all four terminals are higher than the preset threshold, indicating that the fault traveling wave data in this simulation can all be used as valid positioning inputs. Terminal T3 has the highest score, indicating clearer wavefront characteristics and lower signal interference. Terminal N has a relatively lower score, but still meets the threshold requirements. This result demonstrates that the proposed method introduces a data quality assessment mechanism before positioning calculation, which can filter low-quality waveforms and reduce the impact of noise and misidentification on the positioning results.
[0026] The wavefront identification confidence score and the signal quality score are weighted and fused to obtain a comprehensive data quality score; Traveling wave data with a comprehensive data quality score higher than a preset quality threshold are identified as faulty traveling wave data that have undergone quality assessment.
[0027] The preliminary location results of the power distribution line are obtained by parallel calculation using the two-end traveling wave time difference method, the multi-end traveling wave positioning method, and the single-end traveling wave positioning method. The process is as follows: Based on the waveform of the same fault event in the fault traveling wave data that has been quality assessed, three localization algorithms are activated simultaneously; among them, the double-end traveling wave time difference method uses the difference in arrival time of the wavefront of the measurement terminals at both ends of the line and the line length and traveling wave speed to calculate the first fault distance; like Figure 8As shown, a fault occurs outside the section between points A and C on the power line. Traveling waves at two measurement points are captured, and the arrival times t of the two traveling waves are calculated. A and t C If the length of the line is known, the wave speed can be calculated as v = L / (t). A -t C ).
[0028] The multi-terminal traveling wave positioning method uses the combination of wavefront arrival times of three or more measurement terminals to establish a system of simultaneous equations to solve for the second fault distance; The single-ended traveling wave location method calculates the third fault distance using the time delay difference between the reflected wave at the fault point and the initial traveling wave surge. The first fault distance, the second fault distance, and the third fault distance are used together as the preliminary location result.
[0029] Figure 9 The results of fault distance estimation obtained by the two-end traveling wave time difference method, the multi-end traveling wave location method, and the single-end traveling wave location method are presented. Figure 9 The scatter plot represents the localization results output by different algorithms, the error bars represent the uncertainty range of each algorithm's localization result, the red horizontal line represents the mean of the localization results from multiple algorithms, and the light blue area represents the fused localization interval. From Figure 9 As can be seen, the fault distances obtained by the three algorithms are distributed around 6.14 km, 6.20 km, and 6.28 km, respectively. Although there are some deviations between the different algorithms due to their different computational mechanisms, the overall distribution is relatively concentrated, the coefficient of variation is small, and the algorithm consistency score is high. This result indicates that there is a good cross-validation relationship between the two-end time difference method, the multi-end simultaneous solution, and the single-end reflected wave positioning results. By statistically analyzing the dispersion between different effective positioning results, we can determine whether the outputs of multiple algorithms support each other. When the results of multiple algorithms are concentrated, it indicates that the positioning results are highly reliable; when the results of different algorithms are significantly dispersed, it suggests that there may be errors in wavefront identification, wave velocity calibration, or terminal synchronization, requiring further correction.
[0030] Multiple measurement terminals deployed along the power distribution line conduct synchronous data acquisition according to a unified time reference, completely capturing the three-phase current traveling wave waveforms generated during line operation. For each set of acquired three-phase current traveling wave waveforms, wavelet transform modulus maxima method is used to detect the traveling wave front. This method performs layered analysis on the complete traveling wave waveform, extracts the characteristic change points of different levels of the waveform in sequence, and locks the position where the signal undergoes a step change within the waveform based on the inherent characteristics of the traveling wave signal abrupt change. Finally, the corresponding traveling wave front is identified from each set of three-phase current traveling wave waveforms.
[0031] A fixed time interval is defined as the preset analysis window, centered on the location of the detected traveling wave front. The signal energy corresponding to the traveling wave front within this window and the total signal energy contained in the entire preset analysis window are statistically analyzed. The two energy values are compared and the ratio is calculated. This ratio is the confidence level of the wave front identification. For the same three-phase current traveling wave waveform, the effective traveling wave signal and the environmental noise signal are distinguished and their power values are statistically analyzed separately. The signal-to-noise ratio is calculated by combining the two types of power values. Then, according to the preset numerical conversion rules, the obtained signal-to-noise ratio is converted into a corresponding score, which is the signal quality score.
[0032] According to the fixed weight allocation rules set in the technical solution, the corresponding weight values are matched for wavehead identification confidence and signal quality score. The original values of the two indicators are multiplied by their respective matched weight values in turn. The results of the two multiplication operations are then summed. The final value obtained after summing is the comprehensive data quality score.
[0033] Based on the on-site operating conditions of the power distribution line and the pre-set fixed quality threshold according to the signal acquisition standards, each traveling wave waveform is compared with the fixed quality threshold. When the value of the comprehensive data quality score is greater than the preset quality threshold, the traveling wave waveform data is identified as fault traveling wave data that has passed the quality assessment. The relevant formula is as follows: Wave head recognition confidence for: ; in, : Wavehead recognition confidence level, with a value range of [0,1]. Wavefront signal energy Noise energy : Wavelet transform coefficients, where For scale parameters, For time indexing, : The arrival time of the wave head : Analyze the window width; Signal-to-noise ratio calculation ; Wherein, SNR: signal-to-noise ratio, unit: decibel (dB). Signal power, unit: watts (W) Noise power, unit: watt (W); Signal quality score for: ; in, Signal quality score, ranging from [0,1], SNR: signal-to-noise ratio; S4: Data Quality Assessment: Overall Data Quality Score ; in, : Overall data quality score, with a value range of [0,1]. Confidence level for wave head recognition Signal quality score, with weight coefficients B and C for each indicator, which are varied based on sample database data, and generally take values of 0.6 and 0.4.
[0034] The arrival time of the wavefront is obtained from the three-phase current traveling wave waveform synchronously collected from multiple measurement terminals along the power distribution line. The complete waveform is analyzed layer by layer using the wavelet transform modulus maxima method to pinpoint the location where the signal undergoes a step change within the waveform. The analysis window width is set to a fixed duration interval based on the inherent characteristics of traveling wave propagation in power distribution lines, historical waveform data collected continuously over a long period of time, and field application experience in the power industry. This is specifically used to limit the time range for signal energy statistics.
[0035] The time index is generated sequentially by the measurement terminal after collecting traveling wave data according to the device's preset fixed sampling frequency. It corresponds to each independent sampling time node within the analysis window and is used to mark the location of the traveling wave data at different times. The scale parameter is a fixed value obtained through repeated testing and calibration based on a massive field-measured traveling wave data sample library. It is the unified analysis scale used when performing layered analysis of the traveling wave waveform using wavelet transform.
[0036] Wavelet transform coefficients are calculated point-by-point from the raw traveling wave data corresponding to each time index within the analysis window, using a layered wavelet transform analysis process with predetermined scaling parameters. This process reflects the characteristic amplitude state of the traveling wave signal at different locations. The wavefront signal energy is analyzed within a defined window centered on the wavefront arrival time. Wavelet transform coefficients for the corresponding region within the window are extracted, and all extracted values are squared and then summed to obtain the final result.
[0037] The wavelet transform coefficients at positions other than the effective signal area of the wavefront within the complete noise energy extraction analysis window are squared one by one and then accumulated to obtain the final result. The signal power is obtained by dividing the effective traveling wave signal area from the three-phase current traveling wave waveform collected by the measurement terminal, statistically analyzing the signal amplitude of all sampling points in the area, and then uniformly calculating it according to the general calculation rules of power signal power.
[0038] Noise power is calculated by segmenting the noise region caused by environmental interference from the same set of three-phase current traveling wave waveforms, statistically analyzing the interference amplitude at all sampling points within this region, and then uniformly calculating it according to the general calculation rules for power signal power. The fixed critical values used in the signal quality scoring process are comprehensively calibrated based on industry standards for evaluating the quality of traveling wave signals in distribution lines and a large amount of field measurement sample data.
[0039] The two weighting coefficients were determined through multiple rounds of simulation testing and field verification based on a long-term accumulated field traveling wave data sample library. Fixed values were assigned based on the relative importance of wavefront identification and signal quality in the data evaluation system. The wavefront identification confidence score was calculated through a statistical comparison of traveling wave signal energy and noise energy. The signal quality score was generated after a standardized conversion process combining the signal-to-noise ratio with a predetermined critical value.
[0040] The computational logic corresponding to the wavefront identification confidence level quantifies the proportion of effective wavefront signals within the preset analysis window in the overall signal, intuitively reflecting the reliability level of the results obtained from traveling wave wavefront identification and providing a core criterion for evaluating traveling wave data quality. The computational logic corresponding to the signal-to-noise ratio quantifies the relative strength of the effective traveling wave signal and environmental noise in the acquired waveform, objectively reflecting the purity of the original traveling wave acquired waveform, and serving as an important basis for evaluating the quality of basic traveling wave acquisition data.
[0041] The computational logic corresponding to the signal quality scoring can convert the signal-to-noise ratio into a standardized score within a unified numerical range, realizing a quantitative rating of the traveling wave signal quality, which facilitates fusion calculations with other data evaluation indicators. The computational logic corresponding to the comprehensive data quality scoring integrates the wavefront identification confidence level and the signal quality score according to preset weighting rules, forming a single overall evaluation result. This completes the comprehensive judgment of the quality of a single traveling wave waveform data set, thereby filtering out faulty traveling wave data that meets the standards for positioning.
[0042] Complete waveforms belonging to the same fault event within the fault traveling wave data that have completed quality assessment are selected. Using this group of waveforms from the same source as the basis for calculation, three types of positioning calculation processes are simultaneously initiated: double-end traveling wave time difference method, multi-end traveling wave positioning method, and single-end traveling wave positioning method.
[0043] The arrival times of the traveling wave fronts recorded by the measurement terminals at the two ends of the power distribution line are extracted. The difference between the two sets of times is calculated. The calculation is carried out in combination with the overall length of the power distribution line and the propagation speed of the traveling wave in the line. The final result obtained in this calculation is the first fault distance.
[0044] The arrival times of traveling wave fronts recorded by three or more measuring terminals on the power distribution line are extracted. The time data of different terminals are combined in pairs and interrelated operational relationships are established. All interrelated operational relationships are solved as a whole. The final result obtained in this solution is the second fault distance.
[0045] Distinguish between the fault point reflected wave and the initial traveling wave surge within the waveform acquired by a single measurement terminal, calculate the time difference between the arrival times of the two types of traveling waves at the terminal, and perform corresponding calculations based on this time difference. The final result obtained from this calculation is the third fault distance.
[0046] The first fault distance, second fault distance, and third fault distance generated by the calculation are integrated into the same set of data. The integrated distance data is the preliminary location result of the power distribution line.
[0047] The beneficial effects are that the process can complete the synchronous acquisition of traveling wave signals of power distribution lines and the identification of traveling wave fronts, complete data quality verification through multi-level indicators, screen out fault traveling wave data with high reliability, and simultaneously carry out calculations using multiple positioning methods to enrich the reference basis for fault positioning, making the preliminary fault positioning work of power distribution lines more rigorous and further improving the reliability of the positioning results.
[0048] Step 2: Perform statistical analysis on the standard deviation or coefficient of variation of the dispersion between the preliminary positioning results to obtain the algorithm consistency score of the effective positioning results; In this embodiment of the invention, the process of statistically analyzing the standard deviation or coefficient of variation of the dispersion between the preliminary positioning results to obtain the algorithm consistency score of the effective positioning results is as follows: Based on at least two valid localization results obtained from parallel computing, calculate the arithmetic mean of all valid localization results; The standard deviation of the valid positioning results is generated based on the arithmetic mean and each valid positioning result; Multiply the ratio of the standard deviation to the arithmetic mean by 100% to obtain the coefficient of variation of the effective localization results; The coefficient of variation is compared with a preset consistency threshold, and an algorithm consistency score is generated for the effective positioning result according to the rule that the smaller the coefficient of variation, the higher the consistency score.
[0049] The formula for calculating the consistency score of the algorithm is as follows: ; In the formula, The algorithm consistency score is given, with a value range of [0,1]. It is a function with maximum value. The coefficient of variation is 1. This is the consistency threshold.
[0050] For location using the two-end traveling wave time difference method, let the two ends of the line be M and N, the distance from the fault point to end M be x, the line length be L, the wave velocity be v, and the arrival time difference of the wavefronts at both ends be Δt. Then: ; in, The distance from the fault point to end M. Total length of the line Traveling wave propagation speed : The time difference between the arrival of the wavefronts at both ends, Δt=tN-tM, tM: the arrival time of the wavefront at end M, tN: the arrival time of the wavefront at end N; Parallel localization computation using multiple algorithms: Multi-terminal localization, for n terminals (n≥3), establish a system of equations: ; ; ; Where n: number of terminals, n≥3 The fault location calculated by the i-th terminal; Parallel localization computation using multiple algorithms: single-end localization. ; in, : Localization results using the single-ended method; The coefficient of variation is calculated by dividing the standard deviation of the effective positioning results by the arithmetic mean of the effective positioning results and converting it into a percentage. This index is used to reflect the relative dispersion of multiple sets of positioning results.
[0051] The consistency threshold is uniformly set based on the field operation standards for traveling wave positioning of power distribution lines, a large number of test cases, and industry technical specifications, and serves as a fixed reference standard for judging the consistency of the algorithm.
[0052] The algorithm consistency score is generated by comparing the coefficient of variation with the consistency threshold and combining the value range rules. It is used to quantify the degree of uniformity of the output results of different positioning algorithms.
[0053] The maximum value function filters the calculation result and the fixed lower limit value according to the preset value selection rules, and selects the result with the larger value as the final output content, thus constraining the value range of the score.
[0054] The arrival time of the wavefront at the M-end is obtained by a measurement terminal located at one end of the line, which performs wavefront detection on the synchronously acquired three-phase current traveling wave waveform and records the data.
[0055] The arrival time of the N-terminal wavefront is obtained by a measurement terminal located at the other end of the line, which performs wavefront detection on the synchronously acquired three-phase current traveling wave waveform and records the data.
[0056] The arrival time difference of the wavefronts at both ends is obtained by calculating the difference between the arrival times of the wavefronts recorded by the measurement terminals at both ends of the line. It is the basic data for positioning calculations using the two-end traveling wave time difference method.
[0057] The total length of the line is the complete line layout length verified during the construction phase of the power distribution line project, and is continuously verified and confirmed in conjunction with on-site inspections and ledger records.
[0058] The traveling wave propagation speed is determined by combining the conductor material, installation method, operating environment, and long-term historical measurement data of the power distribution line, and represents how fast the transient traveling wave propagates in the line medium.
[0059] The distance from the fault point to end M is calculated based on the total length of the line, the traveling wave propagation speed, and the time difference between the arrival of the wavefronts at both ends, according to the operational logic of the double-end traveling wave time difference method.
[0060] The number of terminals refers to the total number of traveling wave measurement terminals actually deployed and in normal use on the power distribution line. The multi-terminal positioning process can only be executed when this number meets the application condition of three or more terminals.
[0061] In the multi-terminal method, the corresponding segmented line length is divided into sections based on each measurement terminal as a node. The length of each independent section is taken from the official line ledger and verified by on-site measurement.
[0062] In the multi-terminal method, the arrival time difference of the wavefronts at both ends of the corresponding segment is obtained by combining two adjacent sets of measurement terminals, extracting the wavefront arrival times recorded by the terminals respectively, and performing difference calculations.
[0063] The fault location calculated by the i-th terminal is obtained by calculating group by group based on the length of the corresponding segment, the traveling wave propagation speed, and the time difference between the arrival of the wavefronts at both ends of the corresponding segment, according to a unified calculation logic.
[0064] The wavefront time difference used in the single-end method is obtained by distinguishing between the reflected wave from the fault point and the initial traveling wave surge by a single measurement terminal, and then calculating the difference between the arrival times of the two types of traveling waves.
[0065] The single-end method positioning result is calculated based on the traveling wave propagation speed and the wavefront time difference obtained by single-end acquisition, according to the operation rules of the single-end traveling wave positioning method.
[0066] The computational process corresponding to the algorithm consistency score, combined with the coefficient of variation and fixed consistency threshold, completes the conversion and value constraints, which can quantify the discrete state of the output results of various positioning algorithms, intuitively reflect the degree of consistency between the results of different algorithms, and provide a basis for the reliability assessment of positioning results.
[0067] The calculation process corresponding to the double-ended traveling wave time difference method integrates the overall length of the line, the propagation speed of the traveling wave, and the time difference between the wavefronts at both ends of the line. It can accurately calculate the actual distance between the fault location and one end of the line, thus realizing the determination of the basic fault location.
[0068] The complete calculation process corresponding to the multi-terminal method for positioning involves dividing the line section with three or more measurement terminals and performing positioning calculations separately. By using multiple sets of calculation results to form a joint calculation system, the calculation data of the fault location is further enriched, and the reference value of the positioning results is improved.
[0069] The calculation process of the single-end traveling wave positioning method relies on the wavefront time difference and traveling wave propagation speed collected by a single terminal to complete the calculation. It can independently complete the fault distance calculation in scenarios where only single-end terminal data is used, supplementing the diversified positioning results.
[0070] Extract at least two sets of valid positioning results obtained after the parallel positioning operation mentioned above. Sum the values of each valid positioning result sequentially. Divide the sum into equal parts based on the total number of valid positioning results included in the statistics. The value obtained after equal division is the arithmetic mean of all valid positioning results.
[0071] The difference between each valid positioning result and the calculated arithmetic mean is calculated. The difference obtained from each calculation is squared separately. All the squared values are summed together. The sum is then divided equally based on the total number of valid positioning results. The square root of the equalized value is then taken. The result obtained after the entire calculation process is the standard deviation of the valid positioning results.
[0072] Take the calculated standard deviation and arithmetic mean values, divide the standard deviation by the arithmetic mean to get the ratio between the two sets of values, multiply the ratio by one hundred to complete the percentage conversion, and the result after conversion is the coefficient of variation of the effective positioning result.
[0073] By combining the on-site operation specifications and evaluation standards of traveling wave positioning of power distribution lines, a fixed consistency threshold is set in advance as a unified reference benchmark. The calculated coefficient of variation is compared with this fixed reference benchmark. According to the established judgment rule that the lower the value of the coefficient of variation, the higher the value of the final score, the corresponding value is assigned according to the unified score conversion standard, and finally the algorithm consistency score of the effective positioning result is generated.
[0074] The beneficial effects are that this method can objectively calculate the discrete state of different positioning results, quantify the degree of uniformity of the output content of various positioning algorithms, and obtain an algorithm consistency score based on established evaluation criteria, providing a reliable reference for evaluating the overall performance of positioning results.
[0075] Step 3: Based on the error propagation law, evaluate the independent positioning error components of the current power distribution line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error, and obtain the error contribution of each error source and the overall positioning error. In this embodiment of the invention, the independent positioning error components of the current power distribution line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error are evaluated according to the error propagation law to obtain the error contribution of each error source and the overall positioning error. The process is as follows: The estimated values of the current power distribution line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error are obtained. The estimated values are obtained through iterative statistics based on historical fault data. According to the law of error propagation, calculate the positioning error component caused by each error source acting alone: multiply the estimated line length error by one-half to obtain the length error positioning component; Multiply the estimated traveling wave velocity error by half the time difference in the two-end traveling wave time difference method to obtain the wave velocity error positioning component; The synchronization error positioning component is obtained by multiplying the estimated value of the measurement terminal synchronization clock error by half of the traveling wave velocity; the wavefront arrival time error positioning component is obtained by multiplying the estimated value of the wavefront arrival time error by half of the traveling wave velocity. The overall positioning error is calculated using an error synthesis formula, which is: ; In the formula, To account for the overall positioning error, To locate the component of length error, To locate the wave velocity error component, To locate the synchronization error component, The wavefront timing error is located as a component.
[0076] Calculation of the standard deviation of positioning results, assuming k valid positioning results are { If the expression is given by X1, X2, ..., Xk, then the standard deviation σ is: ; ; ; in, The standard deviation of the positioning results reflects the degree of dispersion among algorithms. Number of effective positioning algorithms The localization result of the i-th algorithm. The arithmetic mean of the location results; Location Consistency Analysis: Coefficient of Variation Calculation ; in, Coefficient of variation reflects the relative degree of dispersion. Standard deviation of positioning results; The length error location component is obtained by iteratively statistically analyzing the estimated line length error based on massive historical fault data, and then dividing it equally according to the calculation rules corresponding to the error propagation law.
[0077] The wave velocity error location component is obtained by combining the estimated value of the traveling wave velocity error obtained by iterative statistics based on massive historical fault data with the time difference generated during the operation of the double-ended traveling wave time difference method, and then performing calculations according to the calculation rules corresponding to the error propagation law.
[0078] The synchronization error location component is obtained by calculating the estimated value of the synchronization clock error of the measurement terminal based on the iterative statistics of massive historical fault data, combined with the actual propagation speed of the traveling wave in the power distribution line, and then performing the calculation according to the operation rules corresponding to the error propagation law.
[0079] The wavefront time error positioning component is obtained by calculating the estimated wavefront arrival time error based on the iterative statistics of massive historical fault data, combined with the actual propagation speed of the traveling wave in the power distribution line, and then performing calculations according to the calculation rules corresponding to the error propagation law.
[0080] The comprehensive positioning error calculation process squares the length error positioning component, wave velocity error positioning component, synchronization error positioning component, and wavefront timing error positioning component, respectively. Then, it sums up all the squared values and finally performs a square root operation on the sum. After the entire calculation process is completed, the corresponding result can be obtained.
[0081] The computational logic of this error synthesis strictly follows the law of error propagation. It can summarize the positioning deviations of four different dimensions: line length, traveling wave speed, terminal clock synchronization, and wavefront arrival time, and calculate the overall positioning deviation value generated by the combined effect of multiple errors.
[0082] The comprehensive positioning error obtained from the entire calculation can fully reflect the actual deviation after the sum of all errors in the traveling wave fault location process of the power distribution line, providing accurate and complete basic data for subsequent error contribution analysis, positioning reliability assessment and equipment parameter calibration.
[0083] like Figure 10This illustrates the contribution of four factors—line length error, traveling wave velocity error, synchronization clock error, and wavefront timing error—to the overall positioning error. The horizontal axis represents the error contribution rate, and the vertical axis represents the error source type. Each bar indicates the contribution percentage of its corresponding error source and the resulting positioning error component. In the error propagation analysis method of this invention, each error source is first converted into independent positioning error components, and then the overall positioning error is obtained by taking the square root of the sum of squares. Simultaneously, the contribution rate of each type of error source is determined by the proportion of the square of its error component to the sum of the squares of all error components. Figure 10 As can be seen, the wavefront arrival time error contributes the most and is the dominant error source affecting positioning accuracy in this simulation; followed by traveling wave velocity error, synchronization clock error, and line length error. This indicates that in this scenario, improving wavefront detection accuracy and optimizing the wavefront arrival time identification algorithm are crucial for reducing overall positioning error. These results demonstrate that this method can not only provide the positioning range but also further trace the source of errors, providing direction for on-site parameter correction and algorithm optimization.
[0084] The number of effective positioning algorithms is the total number of positioning algorithms that have performed parallel fault location operations and output valid data. This value is obtained based on the actual operating status statistics of the two-end traveling wave time difference method, the multi-end traveling wave positioning method, and the single-end traveling wave positioning method.
[0085] The location result of the i-th algorithm is the fault distance value calculated by the corresponding location algorithm based on the fault traveling wave data that has undergone quality assessment, according to its own predetermined calculation process.
[0086] The arithmetic mean of the positioning results is obtained by summing the positioning results corresponding to all effective positioning algorithms in sequence, and then dividing the sum equally by combining the number of effective positioning algorithms.
[0087] The standard deviation of the positioning results is obtained by first calculating the difference between each positioning result and the arithmetic mean, squaring each difference, summing all the squared values, dividing the sum equally based on the number of effective positioning algorithms, and finally taking the square root of the average result.
[0088] The coefficient of variation is obtained by comparing the standard deviation of the positioning results with the arithmetic mean of the positioning results, and then converting the result into a percentage.
[0089] The complete calculation process corresponding to the standard deviation of the positioning results can quantify the absolute deviation between the output results of different positioning algorithms and intuitively show the discrete state of the calculation results of various positioning algorithms.
[0090] The calculation process corresponding to the coefficient of variation combines the standard deviation and the arithmetic mean to carry out the conversion work, which can eliminate the influence of the numerical base, objectively reflect the relative dispersion of multiple sets of positioning results, and provide basic data for subsequent determination of the consistency of the algorithm.
[0091] Based on the massive historical fault data stored in the power distribution lines over a long period of time, multiple rounds of iterative statistical operations were carried out in accordance with unified data statistical standards. The estimated values of line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wave front arrival time error were then sorted and summarized in sequence.
[0092] The error propagation rule corresponding to the error propagation law is used to calculate and process the estimated value of the line length error. The complete value corresponding to the estimated value of the line length error is divided equally, and the value obtained after equal division is taken as the length error positioning component.
[0093] The time difference data generated during the execution of the two-end traveling wave time difference method is retrieved, and the traveling wave velocity error estimate is calculated and processed in combination with the error deduction rule corresponding to the error propagation law. The value corresponding to the traveling wave velocity error estimate is calculated and combined with the value obtained after the time difference is evenly divided, and finally the wave velocity error positioning component is formed.
[0094] The velocity value corresponding to the stable propagation of the traveling wave in the power distribution line is retrieved. The error deduction rule corresponding to the error propagation law is used to calculate and process the estimated value of the synchronization clock error of the measurement terminal. The value corresponding to the estimated value of the synchronization clock error of the measurement terminal is calculated and combined with the value obtained after the traveling wave velocity is evenly divided, and finally the synchronization error positioning component is formed.
[0095] The velocity value corresponding to the stable propagation of the traveling wave in the power distribution line is retrieved. The error deduction rule corresponding to the error propagation law is used to calculate and process the error estimate of the wavefront arrival time. The value corresponding to the error estimate of the wavefront arrival time is calculated and combined with the value obtained after the traveling wave velocity is evenly divided, and finally the wavefront time error positioning component is formed.
[0096] Following the multi-source error synthesis operation specification corresponding to the error propagation law, the length error positioning component, wave velocity error positioning component, synchronization error positioning component, and wavefront timing error positioning component are all included in the operation scope. The value of each positioning component is squared separately, all the squared values are accumulated and summarized, and then the overall value is squared. The result obtained after the entire operation process is the comprehensive positioning error.
[0097] The squared results of each independent positioning error component are statistically analyzed separately. All the squared results are added together to obtain a total. The squared result of each individual positioning error component is divided by this total and converted into a percentage to determine the error contribution of each type of error source.
[0098] Step 4: The algorithm consistency score, the comprehensive data quality score of the fault traveling wave data, and the error control confidence obtained from the comprehensive positioning error mapping are weighted and fused to obtain the comprehensive confidence of the effective positioning result; In this embodiment of the invention, the error control confidence level obtained from the comprehensive positioning error mapping is achieved through the following process: The overall positioning error is compared with the preset error control threshold in real time; When the overall positioning error does not exceed the error control threshold, the error control confidence level is calculated according to the ratio of the overall positioning error to the error control threshold, and the error control confidence level decreases monotonically as the overall positioning error increases. When the overall positioning error exceeds the error control threshold, the error control confidence level is set to the minimum value; the range of the error control confidence level is limited to a preset value range.
[0099] The process of weightedly fusing the algorithm consistency score, the comprehensive data quality score of the fault traveling wave data, and the error control confidence score obtained from the comprehensive positioning error mapping to obtain the comprehensive confidence score of the effective positioning result is as follows: Each of the following is assigned a preset weight coefficient: algorithm consistency score, comprehensive data quality score, and error control confidence score. The sum of the weight coefficients is one. The weighted fusion result is obtained by adding the product of the algorithm consistency score and its corresponding weight coefficient, the product of the comprehensive data quality score and its corresponding weight coefficient, and the product of the error control confidence score and its corresponding weight coefficient, and is used as the comprehensive confidence score. The range of the overall confidence level is limited to a preset numerical range, and the higher the value, the higher the reliability of the preliminary positioning result.
[0100] Error source impact quantification analysis: Quantification of the impact of length error, assuming the line length error is... The positioning error caused by the length error is: ; in, Positioning error caused by length error The partial derivative of the positioning result with respect to the line length; Line length error The average error was calculated iteratively based on historical data over the past three months. ; Quantitative Analysis of Error Source Influence: Quantification of the Influence of Wavefront Time Difference Error, assuming the wavefront time difference error is... The positioning error caused by the time error is: ; in, Positioning error caused by wavefront time difference error; Wavehead time difference error The average error was calculated iteratively based on historical data over the past three months. ; Quantitative Analysis of the Impact of Error Sources: Quantification of the Impact of Synchronization Error, assuming the synchronization error is... The positioning error caused by the synchronization error is: ; in, Positioning error caused by synchronization error Time synchronization error; Synchronization error is The average error was calculated iteratively based on historical data over the past three months. ; Quantitative Analysis of Error Source Influence: Quantification of the Influence of Wave Velocity Error, assuming the wave velocity error is... The positioning error caused by the wave velocity error is: ; in, Positioning error caused by wave velocity error; Wave speed error is The average error was calculated iteratively based on historical data over the past three months. ; Quantitative analysis of the impact of error sources: Based on the comprehensive positioning error calculation and using error propagation theory, the total positioning error is: ; in, Total positioning error (overall error) Positioning error caused by length error Positioning error caused by wavefront time difference error. Positioning error caused by synchronization error Positioning error caused by wave velocity error; Confidence Level Assessment: Error Control Confidence Level ; in, Error control confidence level, with a value range of [0,1]. Overall positioning error Error control threshold, typically set at 100 meters; Confidence Level Assessment: Calculation of Overall Confidence Level ; in, Overall confidence level, with a value range of [0,1]. Overall data quality score Algorithm consistency score Error control confidence level, weighting coefficients of each indicator , , The values are typically 0.4, 0.3, and 0.3, with a total sum of 1, based on variations in the sample database. Historical fault and maintenance data of power distribution lines for three consecutive months were extracted, and multiple rounds of iterative statistics were performed according to established statistical standards to finally obtain the average error value corresponding to the line length.
[0101] Based on the correlation between positioning results and line length, and combined with the line length error, the positioning error caused by the length error is calculated.
[0102] Historical data of traveling waves collected from power distribution lines over three consecutive months were extracted. Multiple rounds of iterative statistics were performed according to established statistical standards to finally obtain the average error value corresponding to the wavefront time difference.
[0103] Based on the correlation between the positioning results and the wavefront time difference, and combined with the traveling wave propagation speed and the wavefront time difference error, the positioning error caused by the wavefront time difference error is obtained.
[0104] The historical clock synchronization monitoring data of the measurement terminal for three consecutive months was extracted, and multiple rounds of iterative statistics were performed according to the established statistical specifications to finally obtain the average error value corresponding to time synchronization.
[0105] By combining the traveling wave propagation speed with the time synchronization error, the positioning error caused by the synchronization error is obtained.
[0106] Historical data of traveling wave propagation speed of power distribution lines were extracted over three consecutive months. Multiple rounds of iterative statistics were performed according to established statistical standards to finally obtain the average error value corresponding to the wave speed.
[0107] Based on the correlation between the positioning results and the wave velocity, and combined with the arrival time difference of the wavefronts at both ends and the wave velocity error, the positioning error caused by the wave velocity error is obtained.
[0108] The positioning errors caused by length error, wave head time difference error, synchronization error, and wave velocity error are squared respectively. The square root operation is then performed after summing all the processed values to obtain the comprehensive positioning error.
[0109] An error control threshold is uniformly set based on the requirements for fault location accuracy of power distribution lines, on-site operation and maintenance standards, and industry technical experience. This value serves as a fixed reference for judging the comprehensive location error.
[0110] The comprehensive positioning error is compared and converted with the error control threshold, and then the calculation is completed by combining the preset value constraint rules to obtain the error control confidence level.
[0111] The comprehensive data quality score is a comprehensive evaluation result of the traveling wave data obtained after fusing the wavehead identification confidence score and the signal quality score in the early stage.
[0112] The algorithm consistency score is an evaluation result of the algorithm matching status obtained after statistically analyzing the dispersion of multiple sets of positioning results.
[0113] Based on a large sample library of field measurements, repeated tests and verifications were conducted. Weight coefficients were set according to the importance of the three evaluation indicators in the evaluation system, and all weight coefficients were added together to maintain a fixed value.
[0114] The comprehensive data quality score, algorithm consistency score, and error control confidence score are multiplied by their respective weight coefficients, and the three sets of results are summed to obtain the comprehensive confidence score.
[0115] Error quantification calculations related to line length can accurately calculate the positioning deviation caused by line length deviation alone, and complete the quantitative analysis of this single error source.
[0116] Error quantification calculations related to wavefront time difference can accurately calculate the positioning deviation caused by wavefront time difference deviation alone, and complete the quantitative analysis of this single error source.
[0117] Error quantification calculations related to time synchronization can accurately calculate the positioning deviation caused by the terminal clock synchronization deviation alone, and complete the quantitative analysis of this single error source.
[0118] Error quantification calculations related to traveling wave velocity can accurately calculate the positioning deviation caused by wave velocity deviation alone, and complete the quantitative analysis of this single error source.
[0119] The calculation process for the comprehensive positioning error follows the error propagation theory, integrates four types of independent positioning deviations, and calculates the overall positioning deviation formed by the superposition of various errors.
[0120] The calculation process corresponding to the error control confidence level uses a fixed threshold as a reference standard to convert the comprehensive positioning error into a standardized evaluation score, which intuitively reflects the reliability of the results at the positioning error level.
[0121] The calculation process corresponding to the comprehensive confidence level integrates three types of evaluation indicators according to preset weights to form a unified evaluation result, which fully reflects the comprehensive reliability level of the three dimensions of traveling wave data, positioning algorithm, and positioning error.
[0122] Based on the technical specifications for traveling wave positioning of power distribution lines, on-site fault handling requirements, and measured data accumulated over a long period of operation and maintenance, a fixed error control threshold is pre-set as a unified comparison benchmark. The comprehensive positioning error calculated above is continuously compared with this fixed benchmark.
[0123] When the value of the comprehensive positioning error is within the range of the error control threshold, the proportion of the comprehensive positioning error is calculated using the error control threshold as a reference standard. The score is calculated according to the established conversion rules to obtain the error control confidence level. The value of this indicator will gradually decrease as the value of the comprehensive positioning error increases.
[0124] When the overall positioning error exceeds the fixed value corresponding to the error control threshold, the error control confidence level is directly set to the lowest standard value within the system according to the preset rules of the evaluation system.
[0125] By pre-defining a fixed numerical range in accordance with the unified requirements of the evaluation system for traveling wave location results of power distribution lines, the error control confidence level generated in all scenarios must be within this fixed numerical range, thereby achieving a unified standard for evaluation.
[0126] Based on the weight ratio requirements of the three evaluation indicators—algorithm consistency score, comprehensive data quality score, and error control confidence level—in the overall system, and relying on massive field measurement data and simulation verification results, a unique weight coefficient is pre-assigned to each indicator, and the sum of all weight coefficients remains constant.
[0127] The algorithm consistency score is multiplied by its corresponding weight coefficient, the comprehensive data quality score is multiplied by its corresponding weight coefficient, and the error control confidence score is multiplied by its corresponding weight coefficient. The results of the three calculations are then summed to obtain the comprehensive confidence score.
[0128] In accordance with the unified requirements of the evaluation system for traveling wave location results of power distribution lines, a fixed numerical range is predefined. The calculated comprehensive confidence level must be within this fixed numerical range. The higher the value of this indicator, the higher the reliability of the preliminary location results obtained previously.
[0129] The beneficial effects are that this process can determine the error control confidence level based on the comprehensive positioning error in combination with the established evaluation criteria, and calculate the comprehensive confidence level by integrating multiple evaluation indicators according to reasonable weights. The result rating is completed based on the unified value specification, which intuitively reflects the reliability level of the fault location result, further improves the comprehensive evaluation system of traveling wave location results, and provides a complete and objective reference for on-site personnel to analyze and judge the location data.
[0130] Figure 11This demonstrates the weighted fusion process of three indicators: comprehensive data quality score, algorithm consistency score, and error control confidence score. (The horizontal bar at the top...) Figure 11 The scores for the three evaluation indicators are shown below, and the scale bar below indicates the final overall confidence level and its corresponding level range. In this embodiment of the invention, the overall confidence level consists of three parts: first, an overall data quality score, used to evaluate the reliability of the input traveling wave data; second, an algorithm consistency score, used to evaluate the consistency of results from multiple positioning algorithms; and third, an error control confidence level, used to evaluate whether the overall positioning error is within the allowable range. The three indicators are fused according to preset weights; for example, the overall data quality score has a weight of 0.4, the algorithm consistency score has a weight of 0.3, and the error control confidence level has a weight of 0.3.
[0131] from Figure 11 As can be seen, the overall confidence level obtained from this simulation is approximately 0.81, which falls within the high confidence range. This result indicates that the current fault location results simultaneously meet three conditions: high data quality, good algorithm consistency, and acceptable error control level. Compared to simply outputting the fault distance, this evaluation method can more comprehensively reflect the reliability of the location results, making it easier for on-site personnel to decide whether to directly conduct fault investigation or further verification based on the confidence level.
[0132] Step 5: Determine the confidence level based on the comprehensive confidence level, generate the location range of the fault point in the power distribution line based on the comprehensive location error, and output the identifier of the dominant error source with the largest contribution among all error sources.
[0133] In this embodiment of the invention, the process of determining the confidence level based on the comprehensive confidence level, generating the location range of the fault point in the power distribution line based on the comprehensive location error, and simultaneously outputting the identifier of the dominant error source with the largest contribution among all error sources is as follows: The real-time overall confidence level is compared with multiple preset confidence threshold ranges to determine the confidence level to which the overall confidence level belongs. The confidence levels, from high to low, include at least extremely high confidence, high confidence, medium confidence, low confidence, and insufficient confidence. Using the arithmetic mean of the preliminary positioning results as the positioning reference point, and the comprehensive positioning error as the unilateral offset, the distance of the comprehensive positioning error is offset to the left and right on both sides of the positioning reference point to generate the positioning range of the fault point in the power distribution line. Compare the error contribution of each error source, identify the error source with the largest error contribution as the dominant error source, and output the identification information of the dominant error source.
[0134] The relevant formulas are as follows: Confidence level assessment: Confidence level classification is based on the overall confidence level (Ctotal). ; in, Confidence level, with values ranging from 1 to 2. , , , , , Extremely high confidence level, estimated positioning error <50 meters. High confidence level, estimated positioning error <100 meters. Medium confidence level, estimated positioning error <200 meters. Low confidence level, estimated positioning error <500 meters. The confidence level is insufficient and manual verification is required. Error source contribution analysis and location results The contribution of each error source is as follows: ; ) in, The contribution of the i-th error source, in percentage (%). The positioning error caused by the i-th error source (i=1,2,3,4, corresponding to length, time, synchronization, and wave speed errors, respectively). The positioning error caused by the j-th error source; like If a threshold is set, the corresponding parameters need to be corrected and optimized. Location results:
[0135] in, : Location range :Lower limit of the positioning interval : Upper limit of the positioning interval; The overall confidence score is a comprehensive evaluation index of the positioning results obtained by weighting and fusing the overall data quality score, algorithm consistency score, and error control confidence score according to predetermined weights.
[0136] The confidence level is defined by industry evaluation standards for fault location of power distribution lines and a large amount of on-site measured operation and maintenance data, with fixed levels and corresponding standardized prediction of location error standards matched to different levels.
[0137] The positioning error caused by the i-th error source is a single positioning deviation component obtained by independently calculating each of the four error sources: length error, time error, synchronization error, and wave speed error.
[0138] The positioning error caused by the j-th error source includes the independent positioning deviation components corresponding to all four types of error sources, which are used to complete the summary and statistical calculation of the overall error data.
[0139] The contribution of the i-th error source is obtained by converting the result of the square of a single error component into the sum of the squares of all error components, and is used to characterize the degree of influence of a single error source on the overall positioning deviation.
[0140] The judgment threshold for parameter correction optimization is uniformly calibrated based on the accuracy requirements for fault location and maintenance of power distribution lines and the historical fault measurement data accumulated over a long period of time to form a fixed judgment standard.
[0141] The initial positioning mean is obtained by accumulating and averaging all valid positioning results output by multiple types of traveling wave positioning algorithms, and is the core reference point for constructing the fault location interval.
[0142] The lower limit of the positioning interval is obtained by subtracting the offset distance corresponding to the comprehensive positioning error from the positioning reference point.
[0143] The upper limit of the positioning interval is calculated by superimposing the positioning reference points with the offset distance corresponding to the comprehensive positioning error.
[0144] The location interval is composed of the lower limit and the upper limit of the interval generated by the calculation, and is used to limit the actual distribution range of the fault points of the power distribution line.
[0145] The operational rules for confidence level classification can hierarchically categorize the comprehensive confidence level. Each confidence level corresponds to a fixed positioning error prediction standard, thereby achieving standardized and refined rating of the reliability of fault location results.
[0146] The calculation logic of error source contribution can quantify the proportion of the impact of various error sources on the overall positioning error, accurately distinguish the primary and secondary error sources, and realize comprehensive error source tracing and quantitative analysis.
[0147] The error threshold comparison mechanism can accurately identify error parameters that exceed the allowable deviation range, providing a clear basis for judgment and direction for rectification of power distribution line positioning parameters.
[0148] The logic for constructing the positioning range can combine the positioning benchmark with the overall comprehensive error to generate a fault-tolerant fault range, eliminate the random bias of the positioning results of a single algorithm, and improve the accuracy and practicality of on-site fault diagnosis.
[0149] Based on the evaluation standard for fault location accuracy of power distribution lines and a large amount of field measured sample data, multiple fixed confidence threshold intervals are pre-divided to form a hierarchical confidence rating system. Within the rating system, five fixed levels are set in order of reliability from high to low: extremely high confidence, high confidence, medium confidence, low confidence, and insufficient confidence. The comprehensive confidence obtained in real time is matched one by one with the corresponding threshold interval range to accurately determine the exclusive confidence level corresponding to the current comprehensive confidence.
[0150] All preliminary positioning results output by parallel computation of multiple positioning algorithms are summarized and statistically analyzed. The arithmetic mean of the preliminary positioning results is obtained by accumulating and dividing them equally. This average value is fixed as the reference point for fault location. The calculated comprehensive positioning error is used as a unified unilateral offset scale. With the positioning reference point as the center, the distance ranges offset to both sides of the line by an amount equal to the comprehensive positioning error are finally enclosed to form a complete fault location range for the power distribution line.
[0151] The error contribution data corresponding to the four error sources—line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error—are retrieved. A horizontal comparison and screening process is carried out on all error contribution values. The error source corresponding to the error contribution with the largest value is determined as the dominant error source in the current positioning process. The unique identification information corresponding to the dominant error source is output synchronously to clarify the core error source in this fault location process.
[0152] The beneficial effects are that this method can classify the corresponding confidence level according to the comprehensive confidence level, clearly distinguish the reliability level of the fault location result, define the standard fault location range based on the benchmark point and comprehensive location error, and accurately identify the main error sources affecting the location result and output the corresponding labels. This makes it easier for staff to intuitively judge the credibility of the result and lock the fault range, and also provides clear guidance for subsequent error correction and parameter optimization.
[0153] Figure 12 This simulation illustrates the deployment locations of measurement terminals on the power distribution line, the fault location reference point, and the final generated fault location interval. The horizontal axis represents the line location, and M-terminal, T2, T3, and N-terminal represent the measurement terminals deployed along the line. The red vertical line represents the fused fault location reference point, and the red semi-transparent area represents the fault location interval determined by the comprehensive location error. This embodiment first calculates the arithmetic mean of the effective location results obtained by the dual-end time difference method, multi-end positioning method, and single-end positioning method, using this as the fault location reference point; then, the comprehensive location error is used as a unilateral offset to form a location interval on both sides of the reference point. In this simulation, the location reference point is approximately 6.210 km, and the location interval is approximately [6.170 km, 6.250 km]. Figure 12 This demonstrates the final output format of this method: it not only provides the distance to a single fault but also a location interval with an error tolerance range. Compared to traditional single-point location results, the location interval can better adapt to on-site measurement errors and algorithm uncertainties, helping to narrow the inspection range and improve fault diagnosis efficiency. Furthermore, by combining the aforementioned error contribution and overall confidence results, on-site personnel can simultaneously obtain three key types of information: "possible fault location," "result reliability," and "main error sources."
[0154] As can be seen from the above embodiments, the present invention provides a power distribution line traveling wave location method based on multi-factor analysis. The present invention can quantify the influence weight of four types of factors on the location error: line length, traveling wave velocity, clock synchronization, and wavefront identification time. This enables accurate error tracing, and at the same time, it can evaluate data quality in real time, analyze algorithm consistency, classify location information levels, clarify the reliability of results, and complete parameter correction and error compensation based on error contribution. This effectively reduces the overall location error, significantly improves the fault location accuracy of power distribution lines, and promotes a comprehensive improvement in the efficiency and intelligence level of fault handling.
[0155] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.
[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims
1. A method for locating traveling waves in power distribution lines based on multi-factor analysis, characterized in that, The method includes: Based on the fault traveling wave data collected from the power distribution line after quality assessment, parallel calculations were performed using the double-ended traveling wave time difference method, the multi-ended traveling wave location method, and the single-ended traveling wave location method to obtain the preliminary location results of the power distribution line. Statistical analysis is performed on the standard deviation or coefficient of variation of the dispersion between the preliminary positioning results to obtain the algorithm consistency score of the effective positioning results; Based on the law of error propagation, the independent positioning error components of the current power distribution line's length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error are evaluated respectively. The error contribution of each error source and the overall positioning error are obtained, including: The system obtains the estimated values for the current power distribution line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error. These estimates are obtained through iterative statistical analysis of historical fault data. Based on the error propagation law, the system calculates the location error components caused by each error source acting alone: multiplying the estimated line length error by half yields the length error location component; multiplying the estimated traveling wave velocity error by half the time difference in the two-end traveling wave time difference method yields the wave velocity error location component; multiplying the measurement terminal synchronization clock error by half the traveling wave velocity yields the synchronization error location component; and multiplying the estimated wavefront arrival time error by half the traveling wave velocity yields the wavefront arrival time error location component. Finally, the system calculates the overall location error using an error synthesis formula: ; In the formula, To account for the overall positioning error, To locate the component of length error, To locate the wave velocity error component, To locate the synchronization error component, Positioning components for wavefront timing error; The algorithm consistency score, the comprehensive data quality score of the fault traveling wave data, and the error control confidence obtained from the comprehensive positioning error mapping are weighted and fused to obtain the comprehensive confidence of the effective positioning result. The confidence level is determined based on the overall confidence level, and the location range of the fault point in the power distribution line is generated based on the overall location error. Simultaneously, the identifier of the dominant error source with the largest contribution among all error sources is output. Location results: ; in, : Location range :Lower limit of the positioning interval : Upper limit of the positioning interval : The arithmetic mean of the positioning results.
2. The method for locating traveling waves in power distribution lines based on multi-factor analysis as described in claim 1, characterized in that, The fault traveling wave data collected from the power distribution line and subjected to quality assessment includes: The wavefronts of the three-phase current traveling wave waveforms synchronously acquired by multiple measurement terminals of the power distribution line are detected by wavelet transform modulus maxima method. The confidence level of wavefront identification is calculated based on the ratio of the wavefront energy of the traveling wavefront to the total energy within the preset analysis window. At the same time, the signal-to-noise ratio of the same waveform is calculated and mapped to the signal quality score. The wavefront identification confidence score and the signal quality score are weighted and fused to obtain a comprehensive data quality score; Traveling wave data with a comprehensive data quality score higher than a preset quality threshold are identified as faulty traveling wave data that have undergone quality assessment.
3. The method for locating traveling waves in power distribution lines based on multi-factor analysis as described in claim 2, characterized in that, The method employs parallel calculations using the two-end traveling wave time difference method, the multi-end traveling wave positioning method, and the single-end traveling wave positioning method to obtain preliminary positioning results for the power distribution line, including: Based on the waveform of the same fault event in the fault traveling wave data that has undergone quality assessment, three localization algorithms are simultaneously activated; among them: The two-end traveling wave time difference method calculates the first fault distance by using the difference in arrival time of the wavefronts at the two ends of the line measurement terminals and the line length and traveling wave velocity. The multi-terminal traveling wave positioning method uses the combination of wavefront arrival times of three or more measurement terminals to establish a system of simultaneous equations to solve for the second fault distance; The single-ended traveling wave location method calculates the third fault distance using the time delay difference between the reflected wave at the fault point and the initial traveling wave surge. The first fault distance, the second fault distance, and the third fault distance are used together as the preliminary location result.
4. The method for locating traveling waves in power distribution lines based on multi-factor analysis as described in claim 1, characterized in that, The statistical analysis of the standard deviation or coefficient of variation of the dispersion among the preliminary positioning results to obtain the algorithm consistency score of the effective positioning results includes: Based on at least two valid localization results obtained from parallel computing, calculate the arithmetic mean of all valid localization results; The standard deviation of the valid positioning results is generated based on the arithmetic mean and each valid positioning result; Multiply the ratio of the standard deviation to the arithmetic mean by 100% to obtain the coefficient of variation of the effective localization results; The coefficient of variation is compared with a preset consistency threshold, and an algorithm consistency score is generated for the effective positioning result according to the rule that the smaller the coefficient of variation, the higher the consistency score.
5. The method for locating traveling waves in power distribution lines based on multi-factor analysis as described in claim 4, characterized in that, The formula for calculating the consistency score of the algorithm is as follows: ; In the formula, The algorithm consistency score is given, with a value range of [0,1]. It is a function with maximum value. The coefficient of variation is 1. This is the consistency threshold.
6. The method for locating traveling waves in distribution lines based on multi-factor analysis as described in claim 1, characterized in that, The method of obtaining the error contribution of each error source includes: Calculate the square values of the length error positioning component, wave velocity error positioning component, synchronization error positioning component, and wavefront timing error positioning component respectively; Summing up all the squared values yields the sum of squares. Then, each squared value is calculated as a percentage of the sum of squares. Each percentage is multiplied by 100% to convert it into a corresponding percentage, thus obtaining the contribution of line length error, traveling wave velocity error, measurement terminal synchronization clock error, and wavefront arrival time error. Each error contribution percentage is used to quantitatively identify the dominant error source of the current positioning error, and the error source with the highest contribution percentage is determined as the dominant error source.
7. The method for locating traveling waves in power distribution lines based on multi-factor analysis as described in claim 1, characterized in that, Error control confidence derived from the integrated positioning error mapping includes: The overall positioning error is compared with the preset error control threshold in real time; When the overall positioning error does not exceed the error control threshold, the error control confidence level is calculated according to the ratio of the overall positioning error to the error control threshold, and the error control confidence level decreases monotonically as the overall positioning error increases. When the overall positioning error exceeds the error control threshold, the error control confidence level is set to the minimum value; the range of the error control confidence level is limited to a preset value range.
8. The method for locating traveling waves in power distribution lines based on multi-factor analysis as described in claim 7, characterized in that, The step of weightedly fusing the algorithm consistency score, the comprehensive data quality score of the fault traveling wave data, and the error control confidence score obtained from the comprehensive positioning error mapping to obtain the comprehensive confidence score of the effective positioning result includes: Each of the following is assigned a preset weight coefficient: algorithm consistency score, comprehensive data quality score, and error control confidence score. The sum of the weight coefficients is one. The weighted fusion result is obtained by adding the product of the algorithm consistency score and its corresponding weight coefficient, the product of the comprehensive data quality score and its corresponding weight coefficient, and the product of the error control confidence score and its corresponding weight coefficient, and is used as the comprehensive confidence score. The range of the overall confidence level is limited to a preset numerical range, and the higher the value, the higher the reliability of the preliminary positioning result.
9. The method for locating traveling waves in power distribution lines based on multi-factor analysis as described in claim 1, characterized in that, The process involves determining the confidence level based on the comprehensive confidence level, generating the location range of the fault point in the power distribution line based on the comprehensive location error, and simultaneously outputting the identifier of the dominant error source with the largest contribution among all error sources, including: The real-time overall confidence level is compared with multiple preset confidence threshold ranges to determine the confidence level to which the overall confidence level belongs. The confidence levels, from high to low, include at least extremely high confidence, high confidence, medium confidence, low confidence, and insufficient confidence. Using the arithmetic mean of the preliminary positioning results as the positioning reference point, and the comprehensive positioning error as the unilateral offset, the distance of the comprehensive positioning error is offset to the left and right on both sides of the positioning reference point to generate the positioning range of the fault point in the power distribution line. Compare the error contribution of each error source, identify the error source with the largest error contribution as the dominant error source, and output the identification information of the dominant error source.
Citation Information
Patent Citations
Power transmission line fault monitoring method and system based on distributed traveling wave positioning
CN116773958A
Power distribution network fault positioning method based on traveling wave detection
CN120294506A
