A power distribution line tower leakage detection method and system

Abstract

The application discloses a power distribution line tower leakage detection method and system, and relates to the technical field of power distribution network fault detection.The method comprises the following steps: synchronously collecting three-axis magnetic field data of a target tower at a first height position and a second height position, and generating a first spatial magnetic field time sequence and a second spatial magnetic field time sequence; analyzing the first spatial magnetic field time sequence based on a principal component analysis algorithm, and determining a maximum energy principal axis direction vector; determining a projection plane that is orthogonal to the maximum energy principal axis direction vector, and projecting to determine a first orthogonal magnetic field sequence and a second orthogonal magnetic field sequence; performing a weighted spatial difference operation on the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence, and obtaining a leakage magnetic field difference sequence; and performing feature analysis on the leakage magnetic field difference sequence, extracting a leakage feature parameter, and determining a leakage state of the target tower. The application can reduce the misjudgment rate of leakage detection in a complex environment.

Description

Technical Field

[0001] This application relates to the field of power distribution network fault detection technology, and in particular to a method and system for detecting leakage current in power distribution line towers. Background Technology

[0002] The power distribution network is a crucial component of the power system, with its overhead lines widely supported by metal or reinforced concrete poles. During long-term operation, factors such as insulator aging, flashover from pollution, or lightning strikes can easily cause insulation breakdown in these lines. This leads to current flowing along the poles into the ground, creating leakage current. This not only results in energy loss but also seriously threatens the lives of pedestrians and maintenance personnel. Therefore, accurate and real-time detection of pole leakage is a critical aspect of the safe operation and maintenance of the power distribution network.

[0003] In related technologies, the pole leakage current detection method uses magnetic field absolute amplitude determination technology to achieve leakage current monitoring. This method involves fixing a magnetic field sensor on the pole surface to continuously collect absolute amplitude data of the power frequency magnetic field around the pole; smoothing the collected magnetic field amplitude data through a low-pass filter to extract a stable power frequency magnetic field baseline; finally, directly comparing this baseline with a pre-set fixed safety threshold. When the detected magnetic field amplitude exceeds this threshold, the system triggers a leakage current alarm signal, thus completing the monitoring of the pole leakage current status.

[0004] However, the 100-ampere load current flowing normally through the power distribution line also generates a strong power frequency magnetic field, and this load current fluctuates drastically with the user's electricity demand. The magnetic field changes generated by this fluctuation are superimposed on the absolute amplitude data collected by the sensor. When the load current surges or large vehicles pass by nearby, causing spatial magnetic field distortion, the magnetic field amplitude collected by the sensor will rise and exceed the preset fixed safety threshold, triggering a false alarm. The related tower leakage detection method has a high misjudgment rate. Summary of the Invention

[0005] This application provides a method and system for detecting leakage current in power distribution line towers, which can reduce the false judgment rate of leakage current detection in complex environments.

[0006] In a first aspect, this application provides a method for detecting leakage current in power distribution line towers, applied to a leakage current monitoring system. The method includes: synchronously acquiring triaxial magnetic field data of the target tower at a first height position and a second height position, generating a first spatial magnetic field time series and a second spatial magnetic field time series respectively; analyzing the first spatial magnetic field time series based on principal component analysis to determine the direction vector of the maximum energy principal axis; determining a projection plane orthogonal to the direction vector of the maximum energy principal axis, and projecting the first and second spatial magnetic field time series onto the projection plane respectively to generate a first orthogonal magnetic field sequence and a second orthogonal magnetic field sequence; performing a weighted spatial difference operation on the first and second orthogonal magnetic field sequences to obtain a leakage electromagnetic field difference sequence; and performing feature analysis on the leakage electromagnetic field difference sequence to extract leakage current characteristic parameters to determine the leakage current state of the target tower.

[0007] In the above embodiments, the leakage current monitoring system synchronously collects magnetic field data at two different heights, uses principal component analysis to identify and separate the main magnetic field interference direction generated by the load current, and then uses projection and weighted spatial difference operation to maximize the cancellation of the load current magnetic field and other far-field common-mode interference; transforming the detection object from a mixed magnetic field to a differentiated leakage electromagnetic field, significantly improving the signal-to-noise ratio, effectively suppressing background interference, and reducing the false judgment rate of leakage current detection.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the step of analyzing the first spatial magnetic field time series based on the principal component analysis algorithm to determine the direction vector of the maximum energy principal axis specifically includes: segmenting the first spatial magnetic field time series into windows of integer multiples of the power frequency period to obtain a triaxial magnetic field sample set within the current analysis window; constructing a three-dimensional spatial covariance matrix after performing mean removal processing on the triaxial magnetic field sample set; performing eigenvalue decomposition on the three-dimensional spatial covariance matrix to determine the eigenvector corresponding to the maximum eigenvalue as the direction vector of the maximum energy principal axis; calculating the ratio of the maximum eigenvalue to the sum of all eigenvalues ​​as the principal axis energy concentration index, and discarding the triaxial magnetic field data and triggering the data acquisition and analysis of the next window when the principal axis energy concentration index is lower than a preset concentration threshold.

[0009] In the above embodiments, the leakage current monitoring system can accurately extract the direction of the most concentrated energy, i.e. the main direction of the load current magnetic field, from the data itself by segmenting the data, constructing the covariance matrix, and performing eigenvalue decomposition. At the same time, it ensures the validity and reliability of the determined main energy axis direction vector, avoids subsequent analysis based on erroneous or contaminated data, and improves the stability of the algorithm.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, after performing eigenvalue decomposition on the three-dimensional spatial covariance matrix and determining the eigenvector corresponding to the largest eigenvalue as the maximum energy principal axis direction vector, the method further includes: calculating the absolute value of the cosine of the angle between the maximum energy principal axis direction vector obtained in the current window and the historical principal axis direction vector determined in the previous window; when the absolute value of the cosine is lower than a preset direction consistency threshold, replacing the maximum energy principal axis direction vector of the current window with the historical principal axis direction vector; when the absolute value of the cosine is not lower than the preset direction consistency threshold, performing a weighted average and normalizing on the maximum energy principal axis direction vector of the current window and the historical principal axis direction vector to obtain a smoothed principal axis direction vector, which is used as the current maximum energy principal axis direction vector.

[0011] In the above embodiments, the leakage current monitoring system introduces historical principal axis direction vectors for consistency verification and smoothing. It utilizes the physical characteristic that the direction of the load current magnetic field is stable in a short period of time. By comparing the included angle cosine and weighted averaging, it effectively suppresses drastic changes in the calculated principal axis direction caused by instantaneous interference, ensuring the continuity and stability of the principal axis direction vector. This makes the projection plane used to cancel interference more stable and enhances the robustness of the entire detection method against dynamic environmental interference.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, the step of performing a weighted spatial difference operation on the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence to obtain a leakage electromagnetic field difference sequence specifically includes: obtaining a first spatial distance from a first height position to a power distribution conductor and a second spatial distance from a second height position to a power distribution conductor; determining a first weighting coefficient and a second weighting coefficient based on the first spatial distance and the second spatial distance, such that the load current magnetic field component has equal amplitude at the two height positions after weighting; the ratio of the first weighting coefficient to the second weighting coefficient is equal to the ratio of the first spatial distance to the second spatial distance; multiplying the first orthogonal magnetic field sequence by the first weighting coefficient and the second orthogonal magnetic field sequence by the second weighting coefficient, and then performing a vector difference operation point by point at corresponding times to obtain the leakage electromagnetic field difference sequence.

[0013] In the above embodiments, the leakage current monitoring system utilizes the physical law that magnetic field strength is inversely proportional to distance. By setting a weighting coefficient that is proportional to the spatial distance, the magnetic field sequences at two heights are corrected so that the magnetic field components originating from the load current of the same overhead conductor have equal amplitudes before the differential operation. This allows them to be accurately canceled in the subsequent vector difference operation, improving the suppression effect on the load current magnetic field. As a result, the final differential sequence can more purely reflect the near-field magnetic field generated by the leakage current.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, after determining the first weighting coefficient and the second weighting coefficient based on the first spatial distance and the second spatial distance, the method further includes: extracting the fundamental amplitude of the power frequency of the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence respectively, and calculating the measured amplitude ratio; determining the theoretical amplitude ratio based on the first weighting coefficient and the second weighting coefficient, and calculating the amplitude ratio deviation between the measured amplitude ratio and the theoretical amplitude ratio; when the amplitude ratio deviation exceeds a preset deviation threshold, recalculating the leakage electromagnetic field differential sequence by replacing the theoretical amplitude ratio with the measured amplitude ratio.

[0015] In the above embodiments, the leakage current monitoring system calculates the actual measured magnetic field amplitude ratio and compares it with the theoretical amplitude ratio calculated based on distance. When the deviation is too large, the measured amplitude ratio is used to dynamically adjust the weight, replacing the fixed theoretical value, which compensates for the deficiencies of the theoretical model and the uncertainty of the field environment, making the cancellation of the magnetic field of the load current more accurate and reliable.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, the step of performing feature analysis on the differential sequence of leakage electromagnetic fields and extracting leakage characteristic parameters to determine the leakage state of the target tower specifically includes: performing a Fourier transform on the differential sequence of leakage electromagnetic fields to extract the amplitude at the fundamental frequency point of the power frequency as the differential fundamental amplitude feature; calculating the root mean square value of the differential sequence of leakage electromagnetic fields to obtain the differential effective value feature; calculating the ratio of the fundamental energy of the power frequency to the total energy of the entire frequency band to obtain the spectral purity feature; and constructing a multidimensional leakage feature vector using the differential fundamental amplitude feature, the differential effective value feature, and the spectral purity feature to determine the leakage state of the target tower.

[0017] In the above embodiments, the leakage current monitoring system no longer relies on a single amplitude threshold, but instead constructs a multi-dimensional feature vector to describe the leakage current state; the differential fundamental amplitude, differential effective value, and spectral purity are characterized from three perspectives: the intensity of a specific frequency of the signal, the total energy, and the concentration of frequency components, respectively, which improves the ability to distinguish between real leakage current and various types of interference, making the final judgment result more reliable.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, the step of constructing a multidimensional leakage current feature vector using differential fundamental amplitude features, differential effective value features, and spectral purity features to determine the leakage current state of the target tower specifically includes: obtaining the historical leakage electromagnetic field differential sequence of the target tower under leakage-free operating conditions, and extracting multiple sets of historical multidimensional leakage current feature vectors; calculating the mean and standard deviation of each dimension of the multiple sets of historical multidimensional leakage current feature vectors, and using the mean plus a preset multiple of the standard deviation as the standard detection threshold for the corresponding dimension; determining that the target tower is in a leakage current state when the differential fundamental amplitude features, differential effective value features, and spectral purity features simultaneously exceed their respective standard detection thresholds.

[0019] In the above embodiments, the leakage current monitoring system learns the background noise characteristics of the target tower under normal leakage-free conditions and establishes personalized, statistically based detection thresholds for each feature dimension. This allows it to adapt to the differences between different towers and different electromagnetic environments, reducing false alarms caused by accidental fluctuations in a certain feature and ensuring high confidence in the detection results.

[0020] In a second aspect, embodiments of this application provide a leakage current monitoring system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the leakage current monitoring system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a leakage current monitoring system, cause the leakage current monitoring system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a leakage current monitoring system, cause the leakage current monitoring system to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the leakage current monitoring system provided in the second aspect, the computer storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. Due to the adoption of a dual-sensor synchronous acquisition, principal component analysis to determine the interference principal axis, projection and weighted spatial difference technology, it is possible to effectively separate the strong and dynamically changing load current magnetic field from the acquired signal. This effectively solves the problem in the existing technology that the leakage current detection threshold is difficult to set and the false judgment rate is high due to load current fluctuations or external electromagnetic interference. Thus, it achieves high-precision and high-reliability detection of weak leakage current signals.

[0025] 2. By employing a method that determines the weighting coefficients based on the spatial distance between the sensor and the conductor and performs weighted differential on the orthogonal magnetic field sequence, the load current magnetic field component from the overhead conductor can be accurately canceled in the differential operation based on the physical law that the magnetic field decays with distance. This effectively solves the problem in the prior art that it is impossible to distinguish between leakage electromagnetic field and load current magnetic field, thereby achieving strong suppression of background interference.

[0026] 3. By employing a scheme that extracts differential fundamental amplitude, differential effective value, and spectral purity to construct a multi-dimensional feature vector, the unique characteristics of leakage current signals can be comprehensively characterized from multiple dimensions such as amplitude, energy, and frequency domain distribution. This effectively solves the problem in existing technologies where relying solely on a single magnetic field amplitude is susceptible to transient noise interference and misjudgment, thereby achieving accurate and robust determination of leakage current status. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of a scenario for the leakage current detection method of power distribution line towers in this application embodiment; Figure 2 This is a flowchart illustrating a method for detecting leakage current in power distribution line towers in an embodiment of this application. Figure 3 This is another flowchart illustrating the leakage current detection method for power distribution line towers in this application embodiment; Figure 4 This is a schematic diagram of the physical device structure of a leakage current monitoring system in the embodiments of this application. Detailed Implementation

[0028] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0029] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0030] In the implementation scenario of this application, several technical terms are involved. For example, the target tower refers to a metal or reinforced concrete tower in a power distribution network that requires leakage current monitoring. Three-axis magnetic field data refers to the magnetic field intensity components along three orthogonal coordinate axes (e.g., X, Y, Z axes) measured at a point in space using a three-axis magnetic field sensor, which constitute the magnetic field vector at that point. A spatial magnetic field time series refers to a vector sequence formed by continuously collecting the aforementioned three-axis magnetic field data points over a period of time. Principal component analysis (PCA) is a multivariate statistical analysis method used to project multidimensional data into a lower-dimensional space while preserving the characteristic of maximum variance in the data; in this application, it is used to identify the main magnetic field direction generated by the load current. The maximum energy principal axis direction vector is the direction with the largest data variance identified by the algorithm, corresponding to the direction of the most drastic change in the load current magnetic field physically. Weighted spatial difference operation is a signal processing technique that assigns different weights to signals from two spatial locations and then subtracts them to cancel out interference signals from a common source; here, it is used to eliminate the magnetic field influence of overhead conductor load current. Leakage current characteristic parameters are a set of quantitative indicators extracted from differential signals to characterize the leakage current state, such as differential fundamental amplitude and RMS value. The combined application of these techniques and terms constitutes the core logic of this application in solving the leakage current detection problem in complex environments. Its necessity lies in the fact that only through precise signal processing and feature extraction can weak leakage current signals be reliably separated and identified from background interference that far exceeds the strength of the leakage electromagnetic field.

[0031] Please see Figure 1 This is a schematic diagram of a scenario for the leakage current detection method of power distribution line towers in this application embodiment. Figure 1 As shown, the scenario mainly includes the target tower, power distribution conductors, and a leakage current monitoring system installed on the tower. The leakage current monitoring system includes components installed at a first height position (distance from the conductor). ) and second altitude position (distance from the conductor is The system consists of two triaxial magnetic field sensors and a data processing unit. In actual operation, the load current flowing through the power distribution conductors generates strong far-field magnetic field interference; while when the tower experiences insulation breakdown, the leakage current flows along the tower into the ground, generating a near-field leakage electromagnetic field around the tower. The two sensors at different heights synchronously collect triaxial magnetic field data from their respective spaces and transmit the data to the data processing unit of the leakage current monitoring system. The processing unit, by performing principal component analysis, orthogonal projection, and weighted difference operations based on spatial distance, can maximally cancel the common-mode interference magnetic field generated by the far-end load current, thereby accurately extracting the near-field leakage characteristic parameters and achieving reliable determination of the leakage state.

[0032] The method provided in this embodiment is described in detail below. Please refer to [link / reference]. Figure 2This is a flowchart illustrating a method for detecting leakage current in power distribution line towers in this application.

[0033] S201. Synchronously collect the three-axis magnetic field data of the target tower at the first height position and the second height position, and generate the first spatial magnetic field time series and the second spatial magnetic field time series respectively.

[0034] The target tower refers to a utility pole on the power distribution line to be inspected. The first and second height positions are the vertical spatial positions of two triaxial magnetic field sensors installed on the tower. The triaxial magnetic field data represents the instantaneous values ​​of the magnetic field strength in three directions (Bx, By, Bz) in a Cartesian coordinate system. The first and second spatial magnetic field time series are three-dimensional magnetic field vector sequences obtained by continuous sampling from the sensors at the first and second height positions over a period of time, respectively.

[0035] Specifically, the leakage current monitoring system installs magnetic field sensors at two different locations on the top and bottom of the target tower and activates a synchronous acquisition module. Synchronous acquisition ensures that both sensors record magnetic field data at exactly the same time point, which is the basis for subsequent effective differential calculations to eliminate common-mode interference. The acquired data is stored in time-series format, with each time point corresponding to a three-dimensional magnetic field vector. This vector reflects the total magnetic field generated at that moment and location by all magnetic sources (including line load current, tower leakage current, and environmental interference).

[0036] In some embodiments, synchronous data acquisition can be achieved in several ways: Optionally, a master-slave synchronization scheme can be used, where one sensor is designated as the master device and broadcasts a synchronization clock signal to the slave device via a wired connection (e.g., RS-485 bus). Upon receiving the clock signal, the slave device triggers a data acquisition, ensuring strict synchronization between the two devices. Optionally, a synchronization scheme based on Network Time Protocol (NTP) or Precision Time Protocol (PTP) can be used. Each of the two sensor modules is connected to a local area network supporting the corresponding protocol, obtains a standard time from the same time server via the network, and uses this as a reference for data acquisition. It is understood that other methods can also be used to achieve synchronous data acquisition, such as using a GPS module to provide a unified time signal for each sensor; this is not limited here. In some embodiments, there may be a slight frequency deviation between the sampling clocks of the two sensors, leading to cumulative time asynchrony after long-term operation. To address this, the leakage current monitoring system performs a preprocessing calibration step before data processing. The leakage current monitoring system periodically extracts a small segment of data from the two time series and calculates their cross-correlation function. If the two sequences are synchronized, the peak of the cross-correlation function should appear at zero delay; if there is a time offset, the peak will deviate from the center. The leakage monitoring system calculates the time offset based on the peak position and digitally shifts or resamples one of the time series to achieve precise alignment of the data on the time axis.

[0037] It should be noted that although this embodiment uses the first and second height positions as examples for specific illustration, in practical applications, the number of height positions of the sensor is not limited to two; it can be three or more height positions. By deploying sensor arrays at multiple height positions on the target tower and simultaneously collecting multiple sets of triaxial magnetic field data, richer spatial magnetic field gradient information can be obtained, thereby constructing a higher-order weighted spatial difference model and further improving the system's ability to cancel and suppress interference from complex multi-source environments. This application does not make specific limitations in this regard.

[0038] S202. Analyze the time series of the first spatial magnetic field based on the principal component analysis algorithm to determine the direction vector of the principal axis of maximum energy.

[0039] Principal Component Analysis (PCA) is a statistical method used for data dimensionality reduction and feature extraction. The maximum energy principal axis direction vector refers to the direction with the largest variance in the distribution of the first spatial magnetic field time series data points in three-dimensional space. This direction is usually closely related to the direction of the load current of the strongest magnetic field source, i.e., the power distribution line.

[0040] Specifically, the leakage current monitoring system selects the first spatial magnetic field time series collected by one of the sensors (e.g., a sensor installed at the first height position) as the analysis object. Since the magnetic field generated by the load current is much larger than that generated by the leakage current, the main contribution of energy (or variance) in this time series comes from the load current. The leakage current monitoring system applies principal component analysis to process the three-dimensional magnetic field vector sample set in this series and calculates the eigenvectors of its covariance matrix. Among them, the eigenvector corresponding to the largest eigenvalue indicates the direction of the most drastic data change, and the leakage current monitoring system determines it as the direction vector of the principal axis of maximum energy.

[0041] In some embodiments, the determination of the principal axis direction vector of maximum energy can be achieved in several ways: Optionally, a batch processing method can be used, where the leakage current monitoring system caches triaxial magnetic field data for a period of time (e.g., several seconds or minutes), and then performs principal component analysis on the entire data block to calculate a stable principal axis direction at once; Optionally, a sliding window incremental update method can be used, where the leakage current monitoring system maintains a fixed-size sliding window, and when a new data point enters the window, the covariance matrix and its eigenvectors are quickly updated using an incremental PCA algorithm to achieve real-time tracking of the principal axis direction. It is understood that other methods can also be used to achieve this step, such as using singular value decomposition (SVD) instead of eigenvalue decomposition of the covariance matrix, which has better numerical stability; this is not limited here.

[0042] In some embodiments, sudden, non-line-related strong magnetic field interference (e.g., the passage of large vehicles) may occur near the tower. This interference may become the dominant energy axis for a short period, causing the calculated dominant axis direction to deviate from the actual load current direction. To address this, the leakage current monitoring system introduces a direction smoothing and anomaly suppression mechanism. The leakage current monitoring system maintains a weighted average of historical dominant axis directions. When the angle between the newly calculated dominant axis direction and the historical average direction exceeds a preset threshold, the leakage current monitoring system determines it as an anomaly and discards it, continuing to use the historical average direction; if the angle is within the threshold, it is weighted and averaged with the historical average direction to update the historical value.

[0043] S203. Determine a projection plane orthogonal to the direction vector of the principal axis of maximum energy, and project the first spatial magnetic field time series and the second spatial magnetic field time series onto the projection plane respectively to generate the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence.

[0044] Here, the projection plane refers to a two-dimensional plane in three-dimensional space that is perpendicular to the direction vector of the principal axis of maximum energy determined in step S202. The first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence are new two-dimensional vector time sequences obtained by projecting each magnetic field vector in the original first and second spatial magnetic field time sequences onto this plane.

[0045] Specifically, after determining the principal axis direction vector representing the main direction of the magnetic field of the load current, the leakage current monitoring system constructs a plane orthogonal to this vector. The normal vector of this plane is the principal axis direction vector. Then, the leakage current monitoring system traverses each three-dimensional magnetic field vector in the first and second spatial magnetic field time series, and calculates the component of each vector on the orthogonal plane through vector projection operations. The physical significance of this projection operation is that it removes the component along the principal direction of the load current in each magnetic field vector, retaining the component in the plane orthogonal to it. Since the leakage current originates from the tower itself (near-field source), its magnetic field direction differs significantly from that of the distant overhead line (far-field source), so the projected sequence can better highlight the leakage electromagnetic field information.

[0046] In some embodiments, the calculated principal axis direction vector of maximum energy may exhibit slight jitter. This jitter causes the projection plane to oscillate, introducing noise into subsequent differential calculations. To address this, the leakage current monitoring system applies a time-smoothing filter, such as an exponential moving average (EMA) filter, after determining the principal axis direction vector. Instead of directly using the principal axis vector calculated in the current window, the system performs a weighted average with the smoothed principal axis vector from the previous time step to obtain the smoothed principal axis vector for the current time step. This more stable vector is then used to define the projection plane.

[0047] S204. Perform a weighted spatial difference operation on the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence to obtain the leakage electromagnetic field difference sequence.

[0048] The weighted spatial difference operation involves multiplying the first orthogonal magnetic field sequence by a first weighting coefficient, the second orthogonal magnetic field sequence by a second weighting coefficient, and then performing vector subtraction on the two weighted sequences at each corresponding time point. The leakage electromagnetic field difference sequence is the final result sequence obtained from this operation, and ideally, it mainly contains the magnetic field signal generated by leakage.

[0049] Specifically, the purpose of this step in the leakage current monitoring system is to maximize the cancellation of the common-mode component generated by the load current that remains in the two orthogonal magnetic field sequences after projection. Although the projection operation removes the component in the principal direction, due to the complexity of the magnetic field distribution, residual effects of the load current may still exist in the orthogonal plane. According to electromagnetic field theory, the magnetic field strength from the same far-field source (load current) is inversely proportional to the distance to the source. Therefore, the leakage current monitoring system sets corresponding weighting coefficients based on the ratio of the distances from the two sensors to the overhead conductor, equalizes the amplitude of the two sequences, and then subtracts them. Since the leakage current originates from the tower and is close to the two sensors, its magnetic field component will not be completely canceled in the difference and is thus retained.

[0050] In some embodiments, three-phase current imbalance or asymmetrical conductor arrangement may exist, resulting in a non-linear magnetic field generated by the load current, which makes the cancellation effect of the distance ratio-based linear weighted model ineffective. To address this, the leakage current monitoring system can employ an adaptive cancellation algorithm based on least squares. The leakage current monitoring system uses one sequence (such as a second orthogonal magnetic field sequence) as a reference signal and uses an adaptive filter (such as an LMS or RLS filter) to predict the common-mode interference component in another sequence. This predicted value is then subtracted from the predicted sequence to obtain the differential signal. This filter can automatically learn and adapt to complex, nonlinear interference transfer functions, achieving a better interference cancellation effect.

[0051] S205. Perform feature analysis on the leakage electromagnetic field differential sequence and extract leakage characteristic parameters to determine the leakage status of the target tower.

[0052] Feature analysis refers to the use of signal processing techniques to extract numerical values ​​that quantify the key characteristics of the leakage electromagnetic field differential sequence. Leakage characteristic parameters refer to these extracted values, such as the amplitude of the fundamental power frequency and the root mean square value of the signal. Leakage status refers to the final judgment result on whether the target tower currently exhibits leakage, such as "normal" or "leaking".

[0053] Specifically, after the preceding processing, the resulting differential electromagnetic field sequence theoretically mainly contains the leakage current signal and some background noise. Since the leakage current is power frequency alternating current, the magnetic field it generates should also have significant power frequency characteristics. The leakage monitoring system performs frequency domain analysis (such as Fourier transform) on this differential sequence to extract the energy or amplitude at the power frequency (e.g., 50Hz or 60Hz) and its harmonic frequencies. Simultaneously, it also calculates the overall energy index of the signal (e.g., root mean square value). These extracted feature parameters are compared with preset thresholds or through a machine learning model. If the feature parameters are significantly abnormal, the target tower is determined to be in a leakage state.

[0054] In some embodiments, feature analysis and state determination can be implemented in several ways: Optionally, a multi-threshold comparison method is used, where the leakage current monitoring system calculates multiple features of the differential sequence, such as the fundamental amplitude of the power frequency, total harmonic distortion, and root mean square value, and sets an independent alarm threshold for each feature. A leakage current alarm is triggered only when multiple (or all) features simultaneously exceed their respective thresholds. Optionally, a pattern recognition method is used, where the leakage current monitoring system pre-collects a large number of differential sequence samples under no-leakage and different degrees of leakage conditions, extracts features, and trains a classifier (such as a support vector machine (SVM) or neural network) using these labeled feature vectors. During actual operation, the real-time extracted feature vectors are input into the trained classifier, which directly outputs the leakage current state determination result. It is understood that other methods can also be used to implement this step, and this is not limited here.

[0055] In some embodiments, the differential sequence may contain a small amount of pulses or broadband noise unrelated to the power frequency, which could contaminate energy-based features (such as the root mean square value) and lead to misjudgments. To address this, the leakage current monitoring system calculates an additional "spectral purity" feature during feature extraction. This feature can be defined as the proportion of the sum of the power frequency fundamental wave and its lower harmonics to the total signal energy. A true leakage current signal has its energy highly concentrated near the power frequency, resulting in high spectral purity; while broadband noise has its energy distributed across a wide frequency band, resulting in low spectral purity. In the final determination, spectral purity is used as a key rejection criterion: even if the energy characteristic exceeds the limit, if the spectral purity is below the threshold, it is judged as interference rather than leakage current.

[0056] The above embodiments provide a basic flowchart for a method of detecting leakage current in power distribution line towers. However, in practical applications, environmental interference and measurement errors may lead to a decrease in the accuracy of certain steps. To further improve the robustness and accuracy of the detection, the method provided in this embodiment is described in more detail below. Please refer to... Figure 3 This is another flowchart illustrating the leakage current detection method for power distribution line towers in this application.

[0057] S301. Synchronously acquire the three-axis magnetic field data of the target tower at the first height position and the second height position, and generate the first spatial magnetic field time series and the second spatial magnetic field time series respectively.

[0058] Refer to step S201, which will not be repeated here.

[0059] S302. Analyze the time series of the first spatial magnetic field based on the principal component analysis algorithm to determine the direction vector of the principal axis of maximum energy.

[0060] Referring to step S202, in some embodiments, the leakage current monitoring system performs a series of refined data screening and preprocessing operations to ensure the accuracy and stability of principal component analysis. Specifically, the leakage current monitoring system segments the first spatial magnetic field time series into windows of integer multiples of the power frequency period to obtain a triaxial magnetic field sample set within the current analysis window. After performing mean removal processing on the triaxial magnetic field sample set, a three-dimensional spatial covariance matrix is ​​constructed. Eigenvalue decomposition is performed on the three-dimensional spatial covariance matrix, and the eigenvector corresponding to the largest eigenvalue is determined as the principal axis direction vector of the maximum energy. The ratio of the largest eigenvalue to the sum of all eigenvalues ​​is calculated as the principal axis energy concentration index. When the principal axis energy concentration index is lower than a preset concentration threshold, the triaxial magnetic field data is discarded and the data acquisition and analysis of the next window is triggered.

[0061] Here, the power frequency cycle refers to the time required for alternating current to complete one cycle. The window length refers to the time segment of data used for a single analysis. The triaxial magnetic field sample set is all the three-dimensional magnetic field vector data collected within this window. The three-dimensional spatial covariance matrix is ​​a 3x3 matrix used to describe the linear correlation between the three magnetic field components. Eigenvalue decomposition is the process of decomposing this matrix into eigenvalues ​​and eigenvectors. The principal axis energy concentration index is the proportion of the largest eigenvalue to the sum of all eigenvalues, reflecting the degree of energy concentration of the data along the principal axis.

[0062] Specifically, the leakage current monitoring system extracts data in segments that are integer multiples of the power frequency cycle. This ensures that the analysis window contains a complete sine wave, avoiding spectral leakage caused by truncation. Constructing the covariance matrix and performing eigenvalue decomposition is the standard procedure for principal component analysis, aiming to find the direction with the largest data variance. The key to this embodiment is the addition of a verification of the principal axis energy concentration index. Ideally, the load current is the main source of the magnetic field, and its magnetic field energy should be highly concentrated in one direction, resulting in the largest eigenvalue being much larger than the other two. If this index is too low, it indicates the presence of multiple magnetic field sources of comparable strength or strong random noise, making the calculated principal axis direction unreliable. Therefore, by setting a threshold to discard these poor-quality data, subsequent analysis can be prevented from being contaminated, ensuring the robustness of the entire algorithm.

[0063] In some embodiments, this step can be refined in several ways: Optionally, before constructing the covariance matrix, bandpass filtering can be applied to the data to retain only the power frequency and its adjacent frequency bands. This can pre-filter out some high-frequency and low-frequency noise, making the principal component analysis more focused on the power frequency-related magnetic field source. Optionally, when discarding data, a counter for consecutive discard windows can be set. If data from multiple consecutive windows is discarded, the system can trigger an auxiliary alarm for "excessive environmental interference" to alert maintenance personnel. It is understood that other methods can also be used to implement this step, which are not limited here.

[0064] In some embodiments, the calculation of the covariance matrix can be highly sensitive to outliers in the data. A single outlier can cause a significant deviation in the calculated principal axis direction. To address this, leakage current monitoring systems can employ more robust covariance matrix estimation algorithms, such as the Minimum Covariance Determinant (MCD) estimation. This algorithm finds a subset of all samples with the smallest covariance matrix determinant, and then calculates the covariance matrix and mean based on this "clean" subset. This method effectively resists data contamination and yields a more robust estimate of the principal axis direction.

[0065] In some embodiments, the leakage current monitoring system introduces a dynamic smoothing and correction mechanism to ensure the continuity and stability of the principal axis direction vector over time. Specifically, the leakage current monitoring system calculates the absolute value of the cosine of the angle between the maximum energy principal axis direction vector obtained in the current window and the historical principal axis direction vector determined in the previous window. When the absolute value of the cosine is lower than a preset direction consistency threshold, the maximum energy principal axis direction vector of the current window is replaced by the historical principal axis direction vector. When the absolute value of the cosine is not lower than the preset direction consistency threshold, the maximum energy principal axis direction vector of the current window and the historical principal axis direction vector are weighted averaged and normalized to obtain a smoothed principal axis direction vector, which is used as the current maximum energy principal axis direction vector.

[0066] The historical principal axis direction vector is the principal axis direction obtained after smoothing from the previous one or several analysis windows. The absolute value of the cosine of the angle is an indicator of the similarity between the directions of two vectors; the closer the value is to 1, the more consistent the directions. The direction consistency threshold is a preset value close to 1 (e.g., 0.99) used to determine whether there has been a significant abrupt change in direction. Weighted averaging is a smoothing technique that obtains a more stable update value by assigning different weights to historical and current values.

[0067] Specifically, this step is based on the physical assumption that the direction of the load current is relatively stable over a short period of time. The leakage current monitoring system compares the newly calculated principal axis direction with the historical direction. If the absolute value of the cosine of the included angle is very low, it indicates that the direction has changed drastically and discontinuously, which is likely due to a calculation error caused by a strong instantaneous disturbance. Therefore, the leakage current monitoring system chooses to trust the more stable historical value and directly replaces the current value with the historical value. If the direction change is within a reasonable range (the cosine of the included angle is high but less than 1), the leakage current monitoring system considers this a normal slow change and updates the principal axis direction smoothly through a weighted average (e.g., exponential moving average). This process acts like a low-pass filter, filtering out high-frequency jitter in the principal axis direction and ensuring the stability of the projection plane.

[0068] In some embodiments, this step can be implemented in several ways: optionally, a fixed threshold method can be used, as described above, employing a fixed directional consistency threshold; optionally, an adaptive threshold method can be used, where the threshold is dynamically adjusted according to the signal-to-noise ratio. When the signal quality is good and the energy is high, a more stringent (higher) threshold can be set to allow for smaller fluctuations; when the signal is weak, the threshold can be appropriately relaxed to accommodate potentially increased computational uncertainty. It is understood that other methods can also be used to implement this step, and no limitation is made here.

[0069] In some embodiments, line switching or modifications may permanently alter the physical path of the load current. In this case, the new, correct main axis direction will differ significantly from the old historical direction. If the historical direction is consistently used, the system will be unable to adapt to the new operating conditions. To address this, the leakage current monitoring system sets up a continuous deviation counter. When the absolute value of the cosine of the included angle is detected to be below the direction consistency threshold for N consecutive times (e.g., 10 consecutive windows), the system determines that a genuine change in operating conditions has occurred, rather than a transient disturbance. At this point, the system resets the historical main axis direction, directly adopts the newly calculated main axis direction as the new reference, and begins a new round of smooth iteration, thereby achieving rapid adaptation to structural changes.

[0070] S303. Determine a projection plane orthogonal to the direction vector of the principal axis of maximum energy, and project the first spatial magnetic field time series and the second spatial magnetic field time series onto the projection plane respectively to generate the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence.

[0071] Refer to step S203, which will not be repeated here.

[0072] S304. Obtain the first spatial distance from the first height position to the power distribution conductor and the second spatial distance from the second height position to the power distribution conductor.

[0073] The first spatial distance and the second spatial distance refer to the straight-line distance from the spatial location points of the two sensors to the main source of magnetic field interference (i.e., overhead power distribution lines). These distances are the theoretical basis for calculating the weighting coefficients.

[0074] Specifically, the leakage current monitoring system needs to acquire these two distance parameters. These parameters can be measured by maintenance personnel using tools such as laser rangefinders during the system installation and deployment phase, and then manually entered into the leakage current monitoring system's configuration parameters. These two distance values ​​form the basis for determining the theoretical weighting coefficients in subsequent step S305.

[0075] In some embodiments, distance can be acquired in several ways: Optionally, a manual input method can be used, providing a parameter setting entry on the system's initialization interface for technicians to fill in the measurement parameters. and Optionally, a semi-automatic calibration method is used. After confirming there is no leakage current in the line and the load is stable, the system prompts the user to trigger a calibration process. The system records the magnetic field magnitude values ​​measured by the two sensors at this time. and Because B and Proportional, the system can calculate the distance ratio Users only need to enter one of the distances. The system can then automatically calculate another distance. It is understandable that this step can be achieved in other ways, and no specific method is specified here.

[0076] In some embodiments, the power distribution conductor may sway or sag due to wind or temperature changes, causing the instantaneous distance between the sensor and the conductor to change. A fixed distance parameter cannot reflect this dynamic change. To address this, the leakage current monitoring system can introduce an adaptive correction mechanism in subsequent steps (as described in the embodiments after step S305), which indirectly tracks changes in the distance ratio by monitoring changes in the magnetic field amplitude ratio in real time and dynamically adjusts the weights, rather than relying on initial, static distance measurements.

[0077] S305. Determine the first weighting coefficient and the second weighting coefficient based on the first spatial distance and the second spatial distance, so that the load current magnetic field component has the same amplitude at the two height positions after weighting.

[0078] The first and second weighting coefficients are multipliers used in the weighted difference operation in step S306. Theoretically, their ratio should be equal to the ratio of the first spatial distance to the second spatial distance, i.e. .

[0079] Specifically, the leakage current monitoring system uses the distance obtained in step S304... and Calculate the weighting coefficients and The most direct way to set it is to... , where k is any positive constant. To simplify the calculation, we can directly take k=1, that is... = , = For the magnetic field generated by the same load current I from the overhead conductor, the strengths at the two sensor locations are respectively After weighting, the amplitudes of the two signals become ,as well as .

[0080] The equal magnitudes of both create conditions for subsequent differential cancellation.

[0081] In some embodiments, the field environment may be complex, with other minor common-mode interference sources located in different positions in addition to overhead conductors. In such cases, a single weighting model based on conductor distance may not be optimal. To address this, the leakage current monitoring system can employ optimization algorithms during the calibration phase to find the optimal weights. The leakage current monitoring system aims to minimize the energy (squared root mean square value) of the difference sequence and iteratively searches for the optimal ratio of w1 to w2 using algorithms such as gradient descent. This ratio comprehensively considers the influence of all common-mode interference sources, achieving the best interference cancellation effect.

[0082] In some embodiments, the leakage current monitoring system introduces an adaptive weight correction step to address the deviation between the theoretical model and the actual measurement. Specifically, the leakage current monitoring system extracts the fundamental amplitude of the power frequency from the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence, respectively, and calculates the measured amplitude ratio. Based on the first weighting coefficient and the second weighting coefficient, the theoretical amplitude ratio is determined, and the amplitude ratio deviation between the measured amplitude ratio and the theoretical amplitude ratio is calculated. When the amplitude ratio deviation exceeds a preset deviation threshold, the leakage electromagnetic field differential sequence is recalculated using the measured amplitude ratio instead of the theoretical amplitude ratio.

[0083] The measured amplitude ratio refers to the ratio of the fundamental amplitude of the power frequency magnetic field actually measured by the two sensors. The theoretical amplitude ratio is the expected amplitude ratio calculated based on the installation distance between the two sensors. The amplitude ratio deviation is the difference between these two ratios, such as relative error. The preset deviation threshold is a threshold used to determine whether the theoretical model has failed.

[0084] Specifically, the leakage current monitoring system first calculates a set of theoretical weights based on the installation distance and uses them for preliminary differential analysis. Simultaneously, the system extracts the fundamental frequency amplitude from two undifferentiated orthogonal magnetic field sequences in parallel and calculates their measured ratio. Theoretically, this ratio should be equal to the distance ratio. The system compares the measured ratio with the theoretical ratio. If the difference is small, the distance-based weighting model is accurate. If the difference exceeds a threshold (e.g., 10%), the theoretical model is no longer applicable due to installation errors, conductor asymmetry, or other factors. In this case, the system abandons the theoretical weights and uses the measured amplitude ratio to determine new weights (e.g., new weight ratio). = The differential calculation is then performed again using this new weight. This forms a closed-loop feedback, allowing the differential cancellation effect to adapt to actual operating conditions.

[0085] In some embodiments, this step can be implemented in several ways: Optionally, a one-time calibration method can be used, where the calibration process is executed once during system initialization or periodic maintenance to determine an optimal set of fixed weights for long-term use; alternatively, a periodic calibration method can be used, where the system automatically executes the calibration process every certain period of time (e.g., every hour) to adapt to slow changes in the environment. It is understood that other methods can also be used to implement this step, and no limitation is made here.

[0086] In some embodiments, the load current may be extremely small, resulting in a very low signal-to-noise ratio. In such cases, the extracted fundamental frequency amplitude is inaccurate, and the calculated measured amplitude ratio fluctuates greatly. Using this to correct the weighting may introduce errors. To address this, the leakage current monitoring system first assesses the signal strength of the two sensors before performing correction. Only when the fundamental frequency amplitude measured by both sensors is higher than a minimum effective signal threshold does the system consider the currently calculated measured amplitude ratio reliable and initiate the correction process. If the signal strength is insufficient, the system pauses the correction and continues to use the previous set of effective weighting coefficients.

[0087] S306. Multiply the first orthogonal magnetic field sequence by the first weighting coefficient and the second orthogonal magnetic field sequence by the second weighting coefficient, and then perform vector difference operation point by point according to the corresponding time to obtain the leakage electromagnetic field difference sequence.

[0088] This step is the same as the core operation of step S204, but this embodiment adds a closed-loop feedback mechanism that dynamically corrects the weights based on measured data.

[0089] Specifically, the leakage current monitoring system first performs a weighted differential calculation according to the theoretical weighting coefficients determined in step S305. However, to address the discrepancy between the theoretical model and the actual situation, the leakage current monitoring system further analyzes the fundamental amplitude values ​​of the power frequency of the two original orthogonal magnetic field sequences and calculates the measured amplitude ratio. Then, this measured amplitude ratio is compared with the distance-based... , The calculated theoretical amplitude ratios are compared. If the deviation between the two exceeds a preset threshold, it indicates that the theoretical weights are inaccurate. In this case, the leakage current monitoring system will use the measured amplitude ratio to correct the weight coefficients in reverse and re-execute the weighted difference operation to obtain a more accurate leakage electromagnetic field difference sequence.

[0090] In some embodiments, this step of correction and calculation can be implemented in several ways: Optionally, a post-correction method can be used, first calculating a preliminary difference sequence using theoretical weights, then calculating the deviation between the measured amplitude ratio and the theoretical amplitude ratio. If the deviation exceeds the limit, the weights are recalculated based on the deviation, and the difference operation is performed again. Optionally, a real-time adaptive adjustment method can be used, using the weight coefficients as parameters of an adaptive filter, continuously and slowly adjusting the weight coefficients with the goal of minimizing the energy of the difference sequence, so that it automatically tracks and adapts to the slow changes in the field environment. It is understood that other methods can also be used to implement this step, which are not limited here.

[0091] In some embodiments, when leakage occurs, the leakage electromagnetic field component is superimposed on the load current magnetic field, causing contamination of the measured amplitude ratio. If this contaminated amplitude ratio is used to correct the weights, it may actually reduce the effectiveness of canceling the load current. To address this, the leakage monitoring system performs a quick preliminary leakage assessment of the current differential signal before performing weight correction. If the preliminary assessment indicates significant leakage characteristics (e.g., the differential signal energy far exceeds historical normal levels), the leakage monitoring system pauses the adaptive weight correction and locks in the optimal weight coefficients determined in the previous leakage-free state. This ensures that the load current cancellation model is stable and accurate when leakage occurs.

[0092] S307. Perform Fourier transform on the leakage electromagnetic field differential sequence and extract the amplitude at the power frequency fundamental frequency point as the differential fundamental amplitude feature.

[0093] The Fourier transform is a mathematical tool for converting time-domain signals to the frequency domain. The power frequency fundamental frequency refers to the basic operating frequency of a power system, such as 50Hz or 60Hz. The differential fundamental amplitude characteristic refers to the amplitude of the leakage electromagnetic field differential sequence in the frequency domain corresponding to the power frequency fundamental frequency; this characteristic directly reflects the strength of the power frequency leakage signal.

[0094] Specifically, the leakage current monitoring system processes the leakage electromagnetic field differential sequence (a two-dimensional vector time series) obtained in step S306. The system first calculates the magnitude of the differential sequence at each time point to obtain a one-dimensional amplitude time series, and then performs a Fast Fourier Transform (FFT) on this sequence. After the transform, a spectrum is obtained. The leakage current monitoring system locates the frequency point corresponding to the power grid frequency on the spectrum, reads the amplitude at that point, and uses it as the first leakage current characteristic parameter, i.e., the differential fundamental amplitude characteristic.

[0095] In some embodiments, the fundamental amplitude can be extracted in several ways: Optionally, an FFT method can be used to perform an FFT on a window of differential sequence data to find the corresponding power frequency point and read its amplitude; alternatively, a digital lock-in amplifier algorithm can be used to multiply the differential sequence with a locally generated sine and cosine reference signals of the same frequency and perform low-pass filtering, and the two DC components obtained can be synthesized to obtain the amplitude and phase of the fundamental wave. This method has extremely strong frequency selectivity and noise immunity. It is understood that other methods can also be used to implement this step, such as the Goertzel algorithm, which is specifically designed to calculate the DFT component of a single specific frequency point and has a higher computational efficiency than the full FFT.

[0096] In some embodiments, the grid frequency may fluctuate slightly around its nominal value (e.g., 50Hz). If the amplitude is extracted only at a fixed 50Hz frequency, inaccurate amplitude extraction (i.e., spectral leakage) will occur when the actual frequency deviates. To address this, a leakage current monitoring system includes a frequency tracking module. This module can analyze the signal spectrum, for example by finding spectral peaks or using phase-locked loop (PLL) technology, to accurately estimate the actual frequency of the current grid. Then, when extracting the fundamental amplitude, this accurately tracked frequency is used instead of the fixed nominal frequency, thus ensuring the accuracy of amplitude extraction.

[0097] S308. Calculate the root mean square value of the leakage electromagnetic field difference sequence to obtain the characteristics of the difference effective value.

[0098] The root mean square (RMS) value is a statistic characterizing the energy intensity of a signal that varies over time. It is defined as the square root of the mean of the squares of the instantaneous values ​​of the signal. The effective value characteristic of the differential signal is the RMS value of the leakage electromagnetic field differential sequence, which reflects the overall energy level of the differential signal at all frequencies.

[0099] Specifically, after acquiring the differential sequence of the leakage electromagnetic field, the leakage current monitoring system performs calculations on the data points within an analysis window. For N differential vectors within the window, the system first calculates the square of the magnitude of each vector, then calculates the average of these squared values, and finally takes the square root. This calculated differential RMS characteristic is used as the second leakage current characteristic parameter. It is sensitive to all components of the signal (fundamental frequency, harmonics, noise) and reflects the increase in total magnetic field energy caused by the leakage event.

[0100] In some embodiments, the root mean square (RMS) value can be calculated in several ways: Optionally, a discrete calculation method can be used, strictly following the above definition, summing, averaging, and then taking the square root of the modulus squares of all sampling points within the window; alternatively, a moving average method can be used. To improve the real-time performance of the calculation, a sliding window approach can be employed, where whenever a new data point enters the window and an old data point leaves, the new RMS value is quickly calculated by incrementally updating the summation term, without having to perform a complete calculation for the entire window each time. It is understood that other methods can also be used to implement this step, and no limitation is made here.

[0101] In some embodiments, a single or a few outliers with extremely large amplitudes may appear in the differential sequence, possibly caused by strong transient electromagnetic pulse interference. These outliers disproportionately increase the calculated root mean square (RMS) value, potentially leading to misjudgments. To address this, the leakage current monitoring system preprocesses the differential sequence before calculating the RMS value. For example, median filtering or more complex outlier detection algorithms (such as those based on the 3σ principle) are used to identify, remove, or correct these outliers. The RMS value is then calculated on the clean sequence, thereby enhancing the system's ability to characterize the actual leakage current signal energy and reducing sensitivity to pulse interference.

[0102] S309. Calculate the ratio of the fundamental frequency energy to the total energy of the entire frequency band to obtain the spectral purity characteristics.

[0103] The fundamental frequency energy refers to the energy of the signal near the fundamental frequency point, which can usually be approximated by the square of the fundamental amplitude. The total energy across the entire frequency band refers to the total energy of the signal throughout the entire analysis band; according to Passevar's theorem, it is equal to the square of the root mean square value of the time-domain signal. The spectral purity characteristic is the ratio of these two, used to measure the degree to which the signal energy is concentrated at the power frequency.

[0104] Specifically, the leakage current monitoring system utilizes the results calculated in steps S307 and S308. Let the differential fundamental amplitude extracted in step S307 be... Then the fundamental wave energy is approximately: Let the effective value of the difference calculated in step S308 be... The total energy across the entire frequency band is Then, the leakage current monitoring system calculates the ratio. This value, between 0 and 1, is used as the third leakage current characteristic parameter. A pure power frequency leakage current signal has most of its energy concentrated in the fundamental frequency, therefore its spectral purity characteristic value will be very close to 1.

[0105] In some embodiments, the spectral purity characteristics can be calculated in several ways: Optionally, the amplitude square ratio method can be used, as described above, directly using the calculated fundamental amplitude and RMS value; alternatively, the band energy integration method can be used, integrating the energy of the frequency band containing the power frequency fundamental peak (e.g., 50±1Hz) on the FFT spectrum to obtain the fundamental energy, then integrating the energy of the entire spectrum to obtain the total energy, and then calculating the ratio. This method is less sensitive to small frequency shifts and is more robust. It is understood that other methods can also be used to implement this step, which are not limited here.

[0106] In some embodiments, the leakage current itself may contain strong harmonic components (e.g., caused by nonlinear load grounding). In this case, considering only the fundamental energy may underestimate the purity of the leakage signal, as harmonic energy is also an effective component of the leakage current. To address this, the leakage current monitoring system can calculate a ratio of "power frequency family energy" to the total energy. When calculating the numerator, the leakage current monitoring system includes not only the fundamental energy but also the energy of lower harmonics such as the 3rd, 5th, and 7th harmonics. This defined spectral purity characteristic can more accurately distinguish periodic leakage signals originating from the power system (containing fundamental and harmonics) from aperiodic, broadband random noise.

[0107] S310. Construct a multidimensional leakage current feature vector using differential fundamental amplitude characteristics, differential effective value characteristics, and spectral purity characteristics to determine the leakage current state of the target tower.

[0108] The multidimensional leakage current feature vector is a vector composed of three feature parameters extracted in steps S307, S308, and S309, namely V = [differential fundamental amplitude, differential RMS value, spectral purity]. Determining the leakage current state is a decision-making process based on comparing this feature vector with a judgment criterion.

[0109] Specifically, the leakage current monitoring system combines these three features calculated in real time into a three-dimensional feature vector. To establish a reliable judgment standard, the system collects historical data for a period of time under the premise of confirming that the tower is safe and free of leakage, and extracts multiple sets of historical feature vectors. Then, it calculates the statistical mean and standard deviation for each dimension (each feature) of these historical vectors. Based on these statistics, the leakage current monitoring system sets an adaptive detection threshold for each feature, for example... ,in and represents the historical mean and standard deviation of the i-th feature, and k is a preset multiple. During real-time monitoring, the leakage current monitoring system only determines that the target tower has a leakage current when all three components of the current feature vector simultaneously exceed their respective standard detection thresholds.

[0110] In some embodiments, leakage current may develop slowly, gradually increasing from nothing to something. A fixed threshold may be insensitive to the early stages of such gradual leakage. To address this, a leakage current monitoring system can incorporate a trend analysis module. This module not only focuses on whether the instantaneous value of the feature vector exceeds the limit, but also continuously monitors its trend over longer time scales (e.g., hours or days). The growth slope is calculated by performing linear regression or moving average on the time series of each feature. Even if the feature value has not yet reached the alarm threshold, if a continuous and significant upward trend is detected, the system can issue an early warning of "suspected leakage" or "deteriorating condition," providing decision support for preventative maintenance.

[0111] In some embodiments, to improve the accuracy and reliability of leakage current detection, the leakage current monitoring system adopts a strategy of adaptively generating thresholds based on historical data and performing joint judgment based on multiple features. Specifically, the leakage current monitoring system acquires the historical leakage electromagnetic field differential sequence of the target tower under leakage-free conditions and extracts multiple sets of historical multidimensional leakage current feature vectors. The mean and standard deviation of each dimension of the multiple sets of historical multidimensional leakage current feature vectors are calculated, and the mean plus a preset multiple of the standard deviation is used as the standard detection threshold for the corresponding dimension. When the differential fundamental amplitude feature, differential effective value feature, and spectral purity feature all exceed their respective standard detection thresholds, the target tower is determined to be in a leakage current state.

[0112] The historical leakage electromagnetic field differential sequence refers to the differential sequence collected and processed when the tower is confirmed to be normal; it represents the background noise level of the system. The standard detection threshold is a boundary tailored to each feature, distinguishing between normal and abnormal conditions.

[0113] Specifically, the system first enters a learning phase, collecting and analyzing background signal characteristics under normal operating conditions to establish a digital profile of the normal state. This profile is defined by the mean and standard deviation of three features (fundamental amplitude, RMS value, and spectral purity). The resulting threshold is a personalized threshold for a specific tower and installation location, far more accurate than a universally applicable fixed threshold. During the monitoring phase, the system uses strict AND logic for judgment. A real leakage event will simultaneously cause an increase in the power frequency component of the differential signal (fundamental amplitude exceeding the limit), an increase in total energy (RMS value exceeding the limit), and a prominence of the signal's power frequency characteristics (spectral purity exceeding the limit). Many interferences, such as a brief electromagnetic pulse, may only cause the RMS value to exceed the limit, but the fundamental amplitude and spectral purity may not meet the requirements. This multi-feature joint decision-making mechanism improves immunity to interference and reduces the false alarm rate.

[0114] In some embodiments, this step can be implemented in several ways: Optionally, a hierarchical alarm method can be used. When only one or two features exceed the limit, the system may not directly report leakage, but instead output an intermediate state of "caution" or "suspected abnormality." Only when all three features exceed the limit will the highest-level "leakage" alarm be reported. Optionally, historical normal feature vectors can be fitted into an ellipsoid in a three-dimensional feature space, where the ellipsoid represents the distribution range of normal states. When a new feature vector falls outside the ellipsoid, it is determined to be a leakage. It is understood that other methods can also be used to implement this step, and no limitation is made here.

[0115] In some embodiments, the ambient background noise level itself may change over time (e.g., electromagnetic environment deterioration due to factory operation during the day). Using only a fixed threshold from the initial learning phase may result in false alarms when background noise increases. To address this, the leakage current monitoring system can introduce a dynamic threshold update mechanism. When the system determines that there is currently no leakage current (e.g., all feature values ​​are well below the threshold), it continuously and slowly updates the calculation of historical mean and standard deviation with new "normal" feature vectors. In this way, the detection threshold can automatically adapt to long-term, slow changes in the environmental background, maintaining constant detection sensitivity and false alarm rate, as if "floating" above the background noise level.

[0116] In this embodiment, due to the use of a signal processing method based on dual (multi) sensor spatial difference and principal component analysis, as well as an intelligent decision-making mechanism of multi-dimensional feature fusion and adaptive threshold, it is possible to accurately separate and identify weak leakage electromagnetic field signals from strong background interference. This effectively solves the problem that existing technologies rely solely on single magnetic field amplitude detection, which is easily affected by load current fluctuations and environmental electromagnetic interference, resulting in high misjudgment rate and low reliability. Thus, it achieves highly sensitive and reliable online monitoring of leakage faults in power distribution line towers.

[0117] The leakage current monitoring system in the embodiments of this invention is described below from the perspective of hardware processing. Please refer to [link / reference needed]. Figure 4 This is a schematic diagram of the physical device structure of a leakage current monitoring system in an embodiment of this application.

[0118] It should be noted that, Figure 4 The structure of the leakage current monitoring system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0119] like Figure 4As shown, the leakage current monitoring system includes a CPU 401, which can perform various appropriate actions and processes according to a program stored in ROM 402 or a program loaded into RAM 403 from storage section 408, such as executing the methods described in the above embodiments. RAM 403 also stores various programs and data required for system operation. The CPU 401, ROM 402, and RAM 403 are interconnected via bus 404. I / O interface 405 is also connected to bus 404.

[0120] The following components are connected to I / O interface 405: input section 406 including audio input devices, push-button switches, etc.; output section 407 including liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 408 including hard disks, etc.; and communication section 409 including network interface cards such as LAN (Local Area Network) cards, modems, etc. Communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as disks, optical disks, magneto-optical disks, semiconductor memories, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0121] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by CPU 401, it performs the various functions defined in the present invention.

[0122] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0123] Specifically, the leakage current monitoring system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, it implements the leakage current detection method for power distribution line towers provided in the above embodiment.

[0124] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the leakage current monitoring system described in the above embodiments; or it may exist independently and not assembled into the leakage current monitoring system. The storage medium carries one or more computer programs, which, when executed by a processor of the leakage current monitoring system, cause the leakage current monitoring system to implement the power line tower leakage current detection method provided in the above embodiments.

[0125] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0126] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

Claims

1. A method for detecting leakage current in power distribution line towers, characterized in that, The method, applied to a leakage current monitoring system, includes: Simultaneously collect triaxial magnetic field data of the target tower at the first and second height positions, and generate the first spatial magnetic field time series and the second spatial magnetic field time series respectively; The first spatial magnetic field time series was analyzed based on the principal component analysis algorithm to determine the direction vector of the principal axis of maximum energy. A projection plane orthogonal to the direction vector of the principal axis of maximum energy is determined, and the first spatial magnetic field time series and the second spatial magnetic field time series are projected onto the projection plane respectively to generate the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence. A weighted spatial difference operation is performed on the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence to obtain the leakage electromagnetic field difference sequence; Feature analysis is performed on the leakage electromagnetic field differential sequence to extract leakage characteristic parameters and determine the leakage state of the target tower.

2. The method according to claim 1, characterized in that, The step of analyzing the first spatial magnetic field time series based on the principal component analysis algorithm to determine the direction vector of the maximum energy principal axis specifically includes: The first spatial magnetic field time series is segmented into segments according to a window length that is an integer multiple of the power frequency period to obtain a triaxial magnetic field sample set within the current analysis window; After performing mean removal processing on the triaxial magnetic field sample set, a three-dimensional spatial covariance matrix is ​​constructed. Perform eigenvalue decomposition on the three-dimensional spatial covariance matrix, and determine the eigenvector corresponding to the largest eigenvalue as the direction vector of the principal axis of the maximum energy; The ratio of the maximum eigenvalue to the sum of all eigenvalues ​​is calculated as the principal axis energy concentration index. When the principal axis energy concentration index is lower than a preset concentration threshold, the triaxial magnetic field data is discarded and the data acquisition and analysis of the next window is triggered.

3. The method according to claim 2, characterized in that, After the step of performing eigenvalue decomposition on the three-dimensional spatial covariance matrix and determining the eigenvector corresponding to the largest eigenvalue as the direction vector of the principal axis of maximum energy, the method further includes: Calculate the absolute value of the cosine of the angle between the maximum energy principal axis direction vector obtained in the current window and the historical principal axis direction vector determined in the previous window; When the absolute value of the cosine of the included angle is lower than a preset direction consistency threshold, the maximum energy principal axis direction vector of the current window is replaced by the historical principal axis direction vector; When the absolute value of the cosine of the included angle is not lower than the preset direction consistency threshold, the maximum energy principal axis direction vector of the current window and the historical principal axis direction vector are weighted averaged and normalized to obtain a smoothed principal axis direction vector, which is used as the current maximum energy principal axis direction vector.

4. The method according to claim 1, characterized in that, The step of performing a weighted spatial difference operation on the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence to obtain the leakage electromagnetic field difference sequence specifically includes: Obtain the first spatial distance from the first height position to the power distribution conductor and the second spatial distance from the second height position to the power distribution conductor; The first weighting coefficient and the second weighting coefficient are determined based on the first spatial distance and the second spatial distance, so that the magnitude of the load current magnetic field component is equal at the two height positions after weighting; the ratio of the first weighting coefficient to the second weighting coefficient is equal to the ratio of the first spatial distance to the second spatial distance. The first orthogonal magnetic field sequence is multiplied by the first weighting coefficient, and the second orthogonal magnetic field sequence is multiplied by the second weighting coefficient. Then, vector difference operation is performed point by point at the corresponding time to obtain the leakage electromagnetic field difference sequence.

5. The method according to claim 4, characterized in that, After the step of determining the first weighting coefficient and the second weighting coefficient based on the first spatial distance and the second spatial distance, the method further includes: The fundamental frequency amplitude values ​​of the first orthogonal magnetic field sequence and the second orthogonal magnetic field sequence are extracted respectively, and the measured amplitude ratio is calculated. The theoretical amplitude ratio is determined based on the first weighting coefficient and the second weighting coefficient, and the amplitude ratio deviation between the measured amplitude ratio and the theoretical amplitude ratio is calculated. When the amplitude ratio deviation exceeds a preset deviation threshold, the leakage electromagnetic field differential sequence is recalculated using the measured amplitude ratio instead of the theoretical amplitude ratio.

6. The method according to claim 1, characterized in that, The step of performing feature analysis on the leakage electromagnetic field differential sequence and extracting leakage characteristic parameters to determine the leakage state of the target tower specifically includes: Perform a Fourier transform on the leakage electromagnetic field differential sequence and extract the amplitude at the power frequency fundamental frequency point as the differential fundamental amplitude feature; Calculate the root mean square value of the leakage electromagnetic field difference sequence to obtain the effective value characteristics of the difference; The ratio of the fundamental frequency energy to the total energy across the entire frequency band is calculated to obtain the spectral purity characteristics; A multidimensional leakage current feature vector is constructed using the differential fundamental amplitude feature, the differential effective value feature, and the spectral purity feature to determine the leakage current state of the target tower.

7. The method according to claim 6, characterized in that, The step of constructing a multidimensional leakage current feature vector using the differential fundamental amplitude feature, the differential effective value feature, and the spectral purity feature to determine the leakage current state of the target tower specifically includes: Obtain the historical leakage electromagnetic field difference sequence of the target tower under the condition of no leakage current, and extract multiple sets of historical multidimensional leakage current feature vectors; The mean and standard deviation of each dimension of the multiple sets of historical multidimensional leakage current feature vectors are calculated respectively, and the mean plus the standard deviation by a preset multiple is used as the standard detection threshold of the corresponding dimension. When the differential fundamental amplitude feature, the differential effective value feature, and the spectral purity feature all exceed their respective standard detection thresholds, the target tower is determined to be in a leakage state.

8. A leakage current monitoring system, characterized in that, The leakage current monitoring system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the leakage current monitoring system to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the leakage current monitoring system, the leakage current monitoring system performs the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on the leakage current monitoring system, it causes the leakage current monitoring system to perform the method as described in any one of claims 1-7.

Images