A method and system for quickly locating a grounding fault of a cable branch box

Abstract

The application relates to a cable branch box grounding fault rapid positioning method and system, and relates to the technical field of power distribution network fault detection. The method comprises the following steps: acquiring a fault vibration signal captured by a distributed optical fiber sensing device and extracting a time stamp, determining whether there is a waveform superposition state; if there is, extracting a dominant frequency component to construct a phase disturbance sequence; performing selective gain compensation according to a center frequency response curve to obtain a compensated sequence; determining an instantaneous fault type, combining a propagation speed correction parameter, a node delay and a time stamp to calculate a propagation distance to obtain initial positioning coordinates; if the initial coordinate error exceeds the boundary, the correction anchor point is queried according to the fault type, and the cable path is interpolated and corrected by using a multimodal wave speed difference coefficient to output the final positioning coordinates. The application can effectively solve the positioning deviation problem caused by signal superposition and attenuation under a complex topology, and realize high-precision spatial positioning of an instantaneous grounding fault.

Description

Technical Field

[0001] This application belongs to the field of power distribution network fault detection technology, specifically a method and system for rapid location of grounding faults in cable branch boxes. Background Technology

[0002] In power distribution networks, cable distribution boxes, as key nodes connecting main lines and user terminals, play a crucial role in power distribution and protection. With the continuous expansion of urban underground cable networks, monitoring the operational status of cable distribution boxes and their leads has become a critical link in ensuring power supply reliability. In existing technologies, locating cable grounding faults primarily relies on electrical testing methods or traveling wave ranging methods. These conventional approaches typically apply high-voltage pulses or specific frequency signals to the faulty cable, utilizing the wave propagation characteristics within the cable and analyzing the time difference between the reflected waveform or traveling wave reaching different monitoring points to estimate the fault distance. Such methods have become relatively mature in long-distance, single-line cable installations, providing some data support for basic maintenance.

[0003] However, in practical applications of complex cable networks, the accuracy and reliability of fault location remain persistent challenges. Especially in complex topologies such as branch box leads with short distances, multiple branches, and winding paths, existing location technologies often fall short of ideal performance. Due to the superposition effect of signals during multi-path transmission and the attenuation characteristics in complex environments, the extraction and identification of fault features are significantly difficult, leading to substantial deviations in the final spatial location results. Furthermore, for transient grounding faults, existing technologies still have significant limitations in capturing instantaneous fault characteristics and achieving high-precision spatial reconstruction, failing to meet the stringent requirements of modern smart grids for rapid and accurate fault handling. Summary of the Invention

[0004] To address the above problems, this application provides a method and system for rapid location of grounding faults in cable branch boxes. This method solves the technical problem that the extraction and identification of fault features are significantly difficult due to the superposition effect of signals during multi-path transmission and the attenuation characteristics in complex environments. As a result, the final spatial location results are prone to large deviations, and it is difficult to achieve high-precision spatial reconstruction for transient grounding faults.

[0005] To achieve the above objectives, the technical solution adopted in this application is as follows: The first aspect of this application provides a method for rapid location of grounding faults in cable branch boxes. The method includes: acquiring fault vibration signals captured by distributed optical fiber sensing devices deployed along the branch cable; extracting the arrival timestamps of the fault vibration signals; and determining whether the fault vibration signals exhibit waveform superposition caused by multipath reflection interference. If the waveform superposition exists, the fault vibration signals are frequency-domain separated, the dominant frequency components are extracted, and a phase disturbance sequence is constructed. When the amplitude of the phase disturbance sequence is lower than a preset attenuation threshold, a corresponding frequency band is selected based on the center frequency response curve corresponding to the dominant frequency component to selectively compensate the phase disturbance sequence for gain, resulting in a compensated phase disturbance sequence. The instantaneous fault type is determined based on the compensated phase disturbance sequence, and the propagation distance of the vibration wave in the cable path is calculated by combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay, and the arrival timestamps to obtain initial location coordinates. If the error of the evaluated initial location coordinates exceeds a preset boundary, a correction anchor point is queried based on the instantaneous fault type, and the cable path is interpolated and corrected using a multimode wave velocity difference coefficient to output the final location coordinates.

[0006] The second aspect of this application provides a rapid location system for grounding faults in cable branch boxes. The system includes: a signal acquisition and interference determination module, used to acquire fault vibration signals captured by distributed optical fiber sensing devices deployed along the branch cable, extract the arrival timestamp of the fault vibration signals, and determine whether the fault vibration signals exhibit waveform superposition caused by multipath reflection interference; a frequency domain separation and phase extraction module, used to perform frequency domain separation on the fault vibration signals if the waveform superposition exists, extract the dominant frequency components, and construct a phase perturbation sequence; and a selective gain compensation module, used to, when the amplitude of the phase perturbation sequence is lower than a preset attenuation threshold, adjust the signal based on the dominant frequency components. The center frequency response curve corresponding to the pilot frequency component is used to select the corresponding frequency band and selectively gain compensate the phase disturbance sequence to obtain the compensated phase disturbance sequence. The initial positioning module is used to determine the instantaneous fault type based on the compensated phase disturbance sequence, and calculate the propagation distance of the vibration wave in the cable path by combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay, and the arrival timestamp to obtain the initial positioning coordinates. The path interpolation correction module is used to query the correction anchor point according to the instantaneous fault type if the error of the evaluated initial positioning coordinates exceeds the preset boundary, and interpolate the cable path using the multimode wave velocity difference coefficient to correct the error, and output the final positioning coordinates.

[0007] This application provides a method and system for rapid location of grounding faults in cable branch boxes. The scheme captures fault vibration signals and extracts arrival timestamps through distributed optical fiber sensing devices, enabling the system to perceive the transient characteristics of the fault in real time. Based on the determination of waveform superposition caused by multipath reflection interference, it effectively distinguishes between real fault signals and environmental noise and reflection interference. On this basis, it performs frequency domain separation on fault vibration signals with waveform superposition and extracts the dominant frequency components to construct a phase perturbation sequence. Using techniques such as Hilbert transform, it preserves the phase evolution information of the signal, thus laying the foundation for subsequent accurate compensation.

[0008] When the amplitude of the phase disturbance sequence is lower than the preset attenuation threshold, the corresponding frequency band is selected for selective gain compensation based on the center frequency response curve of the dominant frequency component. This corrects the phase distortion caused by medium attenuation, resulting in a high-fidelity compensated phase disturbance sequence, thereby realizing the recovery and enhancement of weak fault characteristics.

[0009] Subsequently, the instantaneous fault type is determined based on the compensated phase disturbance sequence, and the propagation distance of the vibration wave in the cable path is calculated by combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay, and the arrival timestamp to obtain the initial positioning coordinates. This process integrates multiple physical parameters such as temperature, medium characteristics, and topology node delay, effectively eliminating the system errors introduced by environmental and structural factors.

[0010] Finally, if the error of the initial positioning coordinates exceeds the preset boundary, the anchor point is queried and corrected according to the instantaneous fault type, and the cable path is interpolated and corrected using the multimodal wave velocity difference coefficient. The final positioning coordinates are output. Through fault type-driven anchor point matching and the introduction of the multimodal wave velocity difference coefficient, the geometric error caused by the cable path is further corrected.

[0011] This effectively solves the technical problems of significant difficulties in extracting and identifying fault features due to the superposition effect of signals during multi-path transmission and the attenuation characteristics in complex environments, which in turn makes it easy for the final spatial positioning results to have large deviations and makes it difficult to achieve high-precision spatial reconstruction for transient grounding faults. Therefore, it avoids the delay in emergency repairs and the expansion of power outage areas caused by inaccurate positioning, and improves the accuracy of fault detection, the reliability of positioning results and the stability of power grid operation in complex cable network environments.

[0012] The technical solution proposed in this application can be integrated into the fault prediction and health management system of cable branch boxes, serving as a link for rapid fault diagnosis and precise location, and providing accurate spatial location data support for subsequent health status assessment and operation and maintenance decisions. Attached Figure Description

[0013] Figure 1A flowchart of a method for rapid location of grounding faults in a cable branch box provided in this application; Figure 2 A flowchart illustrating the rapid location method for grounding faults in cable branch boxes provided in this application for obtaining fault vibration signals and extracting arrival timestamps; Figure 3 A flowchart for determining whether the fault vibration signal has a waveform superposition state caused by multipath reflection interference in the rapid location method for grounding faults in cable branch boxes provided in this application; Figure 4 The flowchart illustrates the process of extracting the dominant frequency component and constructing a phase perturbation sequence from the fault vibration signal in the rapid location method for grounding faults in cable branch boxes provided in this application. Figure 5 The flowchart illustrates the process of selecting a corresponding frequency band to selectively gain compensate for the phase disturbance sequence in the rapid location method for grounding faults in cable branch boxes provided in this application, thereby obtaining the compensated phase disturbance sequence. Figure 6 A flowchart illustrating the monitoring of nonlinear distortion indicators and recording of compensation gain saturation limit values ​​in the rapid location method for grounding faults in cable branch boxes provided in this application; Figure 7 A flowchart illustrating the method for rapid location of grounding faults in cable branch boxes provided in this application, which determines the instantaneous fault type based on the compensated phase disturbance sequence; Figure 8 A flowchart illustrating the calculation of the vibration wave propagation distance in the cable path using optical fiber medium propagation speed correction parameters in the rapid location method for grounding faults in cable branch boxes provided in this application. Figure 9 A flowchart for evaluating whether the error of the initial positioning coordinates exceeds the preset boundary in the rapid location method for grounding faults in cable branch boxes provided in this application; Figure 10 The flowchart illustrates the method for rapid location of grounding faults in cable branch boxes provided in this application, which uses a multi-mode wave velocity difference coefficient to interpolate and correct the cable path to output the final location coordinates. Figure 11 A block diagram of a rapid location system for grounding faults in a cable branch box provided in this application. Detailed Implementation

[0014] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0015] like Figure 1 The diagram shows a flowchart of a method for rapid location of grounding faults in a cable branch box according to an embodiment of this application. The method includes: Step S110: Obtain the fault vibration signal captured by the distributed optical fiber sensing device deployed along the branch cable, extract the arrival timestamp of the fault vibration signal, and determine whether the fault vibration signal has a waveform superposition state caused by multipath reflection interference.

[0016] Distributed fiber optic sensing devices, serving as the foundation of physical sensing, utilize Rayleigh or Brillouin scattering effects to convert mechanical vibration waves propagating along branch cables into electrical signals, thereby detecting weak vibrations caused by fault discharges. The fault vibration signal is the time-domain waveform data collected when the mechanical shock wave generated by the fault discharge acts on the optical fiber, containing characteristic information such as arrival time, frequency components, and phase changes.

[0017] The arrival timestamp is used to mark the moment when the fault vibration wave arrives at a specific sensing location. The process of obtaining this timestamp can be, for example, by using a variance detection algorithm based on a sliding window. When the signal variance of multiple consecutive sampling points exceeds a preset multiple of the background noise variance, the timestamp of the first point exceeding the limit is marked as the arrival time.

[0018] In another alternative implementation, the process of obtaining the arrival timestamp can also be as follows: Figure 2 The diagram shown is a flowchart of a method for acquiring fault vibration signals and extracting arrival timestamps according to an embodiment of this application. The method further includes: Step S210: Collect continuous backscattered light data stream along the branch cable in real time using a distributed optical fiber sensing device at a preset sampling rate.

[0019] Continuous backscattering data streams characterize the continuous temporal and spatial changes in the phase of Rayleigh scattered light, belonging to high-dimensional time-series data streams. They contain strain rate and acceleration information at various points along the cable. The acquisition process includes: injecting highly coherent light into the sensing fiber using a laser emitting unit in a distributed fiber optic sensing device, and real-time acquisition of the backscattered Rayleigh light signal using a high-sensitivity photodetector. During signal acquisition, the analog-to-digital conversion module performs high-speed data acquisition controlled by a preset sampling rate.

[0020] The preset sampling rate is selected based on the Nyquist sampling theorem and takes into account the typical bandwidth of the fault vibration signal, for example, set to 2000Hz, to ensure that the fault vibration characteristics with frequencies below 1000Hz can be fully reproduced. The acquired continuous backscattered light data stream is cached in a memory buffer in the form of timestamp-indexed data frames for subsequent real-time analysis module reading. The format of this data stream is, for example, a structured binary stream containing light intensity values, phase quadrature components I / Q, and the corresponding sampling time. Based on the above high-frequency acquisition mechanism, microsecond-level signal abrupt changes can be captured, providing data support for subsequent precise positioning.

[0021] Step S230: Set short and long time windows, perform real-time analysis on the continuous backscattering data stream, and calculate the energy ratio of the short and long time windows.

[0022] A sliding window algorithm is used to perform real-time scanning analysis of continuous backscattered light data streams, aiming to quickly identify fault signals with transient impact characteristics from persistent background noise. A short window and a long window are two analysis windows with different time spans. The short window is used to capture instantaneous changes in the signal, and its duration is set, for example, to correspond to 50 sampling points. The long window is used to characterize the average energy level of the signal background, and its duration is set, for example, to correspond to 500 sampling points.

[0023] Calculating the energy ratio of the short-time window to the long-time window is the core logic for determining signal abrupt changes. The specific calculation process includes: calculating the sum of the squares of the signal amplitudes at all sampling points within the short-time window as the short-time energy, and simultaneously calculating the sum of the squares of the signal amplitudes at all sampling points within the long-time window as the long-time energy. Then, the short-time energy is divided by the long-time energy to obtain the energy ratio. This ratio reflects the degree of sudden increase in signal energy relative to the background noise level at the current moment. Compared to traditional fixed threshold triggering methods, the energy ratio-based determination mechanism can effectively eliminate the influence of common-mode interference such as ambient temperature drift and light source intensity fluctuations, ensuring accurate identification of fault events even in low signal-to-noise ratio environments.

[0024] Step S250: When the energy ratio is greater than or equal to the preset trigger threshold, it is determined that the fault vibration signal generated at the moment of the fault is captured, and the absolute time corresponding to the sudden change point of the fault vibration signal is extracted as the arrival timestamp.

[0025] When the calculated energy ratio is greater than or equal to the preset trigger threshold, it is determined that the branch cable has experienced a sudden and severe mechanical vibration, i.e., the fault vibration signal has been captured.

[0026] The preset trigger threshold is used to distinguish between normal background fluctuations and fault impacts. It is determined as follows: when the branch cable is running normally and without faults, the root mean square value of the signal amplitude is calculated as the noise floor characterization quantity within a long time window. After continuously collecting multiple sets of noise floor characterization quantities, their mean and standard deviation are calculated. The sum of the mean and the standard deviation of the preset multiple is set as the preset trigger threshold. The preset multiple is preferably a real number between 3 and 5.

[0027] Once the judgment condition is met, the system immediately locks the signal segment corresponding to the current moment and backtracks to analyze the waveform morphology of that segment in order to accurately extract the starting moment of the fault. This process includes: searching for the point with the largest change in signal slope or amplitude in the data segment near the trigger moment, marking it as the abrupt change point, which represents the precise moment when the vibration wavefront first arrives at a certain position in the sensing fiber.

[0028] Subsequently, the high-precision timestamp corresponding to the mutation point in the data stream is read as the absolute time. This absolute time is typically provided by the GPS timing module or BeiDou timing module built into the distributed fiber optic sensing device, possessing nanosecond-level synchronization accuracy. Finally, this absolute time is recorded as an arrival timestamp, for example, 10:20:30:50 milliseconds. Through this automatic extraction mechanism triggered by energy ratios, the delays and errors of manual identification are eliminated, ensuring that the time reference required for subsequent fault location calculations has extremely high reliability and consistency.

[0029] Through the above steps, the continuous backscattered light data stream collected by the distributed fiber optic sensing device is monitored in real time using a long-short time window energy ratio algorithm. The fault vibration signal is automatically captured the instant the energy ratio exceeds a preset trigger threshold, and the absolute time corresponding to the abrupt change is accurately extracted as the arrival timestamp. This automated triggering mechanism simulates the process of transient signals breaking through background steady-state noise, effectively solving the problem of false triggering or missed triggering caused by noise interference in traditional methods. It provides millisecond-level or even microsecond-level time reference data for subsequent high-precision fault location based on time difference, thereby significantly improving the response speed and location accuracy of the cable branch box grounding fault location system.

[0030] Multipath reflection interference refers to the phenomenon where vibration waves undergo multiple reflections and refractions at branch boxes, joints, or medium interfaces, resulting in the superposition of back-propagating waves in the sensing area. This phenomenon causes distortion of the original signal. The waveform superposition manifests as a widening of the time-domain waveform, the appearance of additional peaks / troughs, and non-monotonic changes in the envelope.

[0031] In the embodiments of this application, such as Figure 3 As shown, the logic for determining whether a waveform superposition state exists specifically includes: Step S310: After filtering out background noise from the fault vibration signal, extract the envelope of the processed signal in the time domain and calculate the time interval between adjacent peaks on the envelope.

[0032] Background noise filtering aims to suppress random high-frequency interference in the environment and improve the signal-to-noise ratio of subsequent feature extraction. This process is based on the difference in frequency domain energy distribution between signal and noise, and reconstructs the fault vibration signal using a specific algorithm to restore the true physical vibration pattern. Specifically, wavelet decomposition algorithms can be used, selecting appropriate wavelet basis functions, such as the Daubechies wavelet, to perform multi-scale decomposition of the signal. A soft threshold function is set to filter out noise coefficients in high-frequency detail components, retaining low-frequency approximate components containing the main fault information. Finally, inverse wavelet transform is used to reconstruct the signal, thus obtaining the clean signal after background noise filtering.

[0033] Based on the acquisition of a clean signal, its envelope in the time domain is further extracted. The envelope describes the changing trend of the signal amplitude profile, reflecting the distribution trajectory of signal energy in the time domain. To avoid errors caused by simple peak detection, this embodiment uses analytic signal construction methods such as Hilbert transform to calculate the envelope. Specifically, the real-valued signal after background noise is filtered out is converted into an analytic signal, and the sequence of the magnitude of this analytic signal changing over time is taken; this sequence constitutes a smooth signal envelope. By constructing the envelope, the high-frequency carrier fluctuations within the signal can be effectively smoothed, highlighting the overall framework of signal energy fluctuations and avoiding false extreme points caused by waveform glitches during subsequent peak detection.

[0034] Based on the aforementioned envelope, the time interval between adjacent peaks is calculated to quantify the temporal density of the signal waveform. The position of each peak is determined by searching for local maxima in the envelope sequence, and the sampling time corresponding to each peak is recorded. Then, the difference between the times of adjacent peaks is calculated. If the time interval between adjacent peaks is small, it indicates that the signal waveform has experienced multiple energy fluctuations within a short period, which is typically a characteristic of multipath reflection superposition.

[0035] Step S330: When the time intervals of a preset number of consecutive waves are all less than the preset time threshold, and the ratio of the amplitude of the secondary wave peak to the amplitude of the main wave peak reaches the preset ratio, it is determined that there is a waveform superposition state.

[0036] A preset time threshold is used as a time criterion to distinguish between normal impact signals and superimposed interference signals. It is determined based on the standard pulse width of the cable fault vibration wave propagating along a single path and the sampling rate of the sensor. Specifically, the preset time threshold can be estimated in advance based on the typical length of the branch cable in the cable branch box and the propagation speed of the fault vibration wave in that cable.

[0037] For example, for branch cables less than 50 meters in length, an initial value of 0.5 to 0.8 times the width of a single vibration pulse is selected as the preset time threshold. This value is then fine-tuned based on the measured noise level on-site to ensure that the interval between adjacent peaks of a normal single pulse is greater than the preset time threshold, while the interval between adjacent peaks in multipath reflection scenarios is less than the preset time threshold. For typical single ground fault vibrations, the main pulse width is usually large; however, the echo superposition caused by multipath reflections makes the pulses closer together.

[0038] If the time intervals of a preset number of consecutive intervals are all less than the preset time threshold, it indicates that the signal exhibits an unusually tight continuous oscillation structure in the time domain, satisfying the time dimension characteristics of waveform superposition.

[0039] To eliminate occasional interference and jitter, the preset quantity can be set to, for example, 3, requiring three consecutive dense intervals. The preset quantity setting must balance detection sensitivity and anti-interference capability: too small a preset quantity may easily misjudge a single noise pulse as a superposition state, while too large a preset quantity will lead to a lag in the response to the actual superimposed signal. Considering the number of reflection paths and signal attenuation characteristics in the cable branch box, setting the preset quantity to 3 effectively filters out occasional spike interference while also promptly detecting continuous dense oscillating signal segments formed by multi-path reflections.

[0040] While satisfying the time dimension characteristics, it is also necessary to verify the energy intensity of the secondary peak. The secondary peak refers to a smaller peak that appears after the main peak, usually formed by reflected or scattered waves. When the amplitude of the secondary peak reaches a preset proportion of the amplitude of the main peak, it means that the subsequent peak has sufficiently high energy and is not a negligible noise floor, but a real superposition component.

[0041] For example, a preset ratio can be set to 60%, meaning that when the ratio of the amplitude of the secondary peak to the amplitude of the primary peak is greater than or equal to 60%, it is considered to constitute effective superposition. The selection of the preset ratio needs to be determined based on the signal-to-noise ratio of typical fault signals and the attenuation characteristics of reflected waves relative to direct waves. For multipath reflection scenarios in cable branch boxes, the amplitude of reflected waves is usually between 50% and 80% of the amplitude of the primary wave. Setting the preset ratio to 60% can effectively distinguish between effective reflected components and background noise.

[0042] By constructing a dual-judgment logic based on both temporal density and energy significance, the robustness of waveform superposition identification can be significantly improved. A single time interval threshold is easily affected by high-frequency noise pulses, leading to false judgments, while a single amplitude ratio threshold may miss weak far-end reflection superposition. Combining these two conditions, waveform superposition is only confirmed when the signal exhibits high-density fluctuations in time, and the amplitude of these fluctuations is sufficiently significant. This provides an accurate triggering basis for subsequent frequency domain separation and phase compensation.

[0043] Step S130: If there is a waveform superposition state, the fault vibration signal is separated in the frequency domain, the dominant frequency component is extracted and a phase perturbation sequence is constructed.

[0044] Frequency domain separation aims to decouple frequency components reflecting the essential characteristics of a fault from aliased signals composed of superimposed waveforms. The dominant frequency component refers to the frequency component with the highest energy proportion in the spectrum that represents the characteristics of the fault source, such as the frequency of an electric arc oscillation. The phase perturbation sequence describes the minute phase shifts of the dominant frequency component over time, sensitively reflecting phase distortion during the propagation of vibration waves.

[0045] The extraction process of the dominant frequency components can be, for example, by using a time-frequency analysis method based on wavelet transform to search for energy maxima in the scale space to determine the dominant frequency components.

[0046] In another alternative implementation, the extraction process of the dominant frequency component can also be as follows: Figure 4 As shown, it specifically includes: Step S410: Apply a window function to the processed signal and then perform a fast Fourier transform to obtain the frequency domain energy distribution.

[0047] The processed signal can refer to the low-frequency fault characteristic component signal retained after background noise filtering in the above-described embodiments, which reflects the true physical form of the fault vibration. To eliminate the spectral leakage effect caused by signal truncation in the frequency domain analysis, a window function is first applied to the processed signal. This window function can be, for example, a Hanning window, which, with its wide main lobe and large side lobe attenuation characteristics, can effectively suppress energy leakage of high-frequency components, thereby ensuring the purity of the frequency domain energy distribution.

[0048] After applying a window function, a Fast Fourier Transform (FFT) is performed on the signal. As an efficient algorithm for converting time-domain signals into frequency-domain signals, the FFT decomposes discrete-time signals into sinusoidal components of different frequencies, accurately revealing the frequency components contained within the signal and the energy intensity of each component, thereby obtaining a frequency-domain energy distribution that reflects how the signal energy changes with frequency.

[0049] This frequency domain energy distribution is not a simple list of frequencies, but a structured data matrix that characterizes the energy projection intensity of fault vibration in different frequency bands; its horizontal axis is frequency, and its vertical axis is amplitude spectral density or energy spectral density, which intuitively shows the dominant energy concentration area of ​​the vibration signal.

[0050] Step S430: Traverse the frequency domain energy distribution, identify frequency bands with an energy proportion exceeding a preset proportion threshold as dominant frequency bands, and extract the corresponding dominant frequency components.

[0051] Based on the frequency domain energy distribution obtained above, in order to pinpoint the core characteristics of the fault source from the complex spectrum containing noise and interference, a comprehensive analysis of the frequency domain energy distribution is required. In this process, a preset proportion threshold serves as a key parameter for distinguishing the primary and secondary positions of signal energy. The specific value of the preset proportion threshold is determined based on the physical characteristics of the fault vibration signal and engineering experience.

[0052] In one implementation, statistical analysis is performed on the frequency domain energy distribution of multiple historical fault samples to calculate the average proportion of the dominant frequency band energy to the total energy. This average proportion is then multiplied by a preset coefficient to obtain a preset proportion threshold. In another implementation, the preset proportion threshold is dynamically adjusted based on the real-time signal-to-noise ratio (SNR): the peak value of the frequency domain energy distribution and the average energy of the background noise are extracted, and the current SNR is estimated based on the ratio of the peak value to the background noise energy. When the SNR is high, for example, greater than 20 dB, the preset proportion threshold is set to a relatively high value, such as 70%, to filter out the most significant dominant frequency components. When the SNR is low, for example, less than 10 dB, the preset proportion threshold is set to a relatively low value, such as 50%, to ensure that effective dominant frequency components can still be extracted even in noisy environments. In a preferred fixed threshold implementation, the preset proportion threshold is set to 60%, meaning that when the energy of a certain frequency band or point accounts for more than 60% of the total energy, that frequency band is identified as the dominant frequency band.

[0053] By statistically analyzing the energy proportion of each frequency band or point in the total energy, frequency bands with an energy proportion exceeding a preset threshold are identified as dominant frequency bands. These dominant frequency bands represent the frequency components in the fault vibration signal that have the most concentrated energy and best reflect the essential characteristics of the fault. Subsequently, based on the identified dominant frequency bands, frequency domain filtering or spectrum extraction algorithms are used to separate the corresponding frequency components from the original spectrum, thereby extracting the dominant frequency components. These dominant frequency components eliminate the interference of secondary frequency components, allowing subsequent phase calculations to focus more on the physical mechanism of the fault itself, thus forming the basic data carrier for subsequent phase analysis.

[0054] In this embodiment of the application, after extracting the dominant frequency components, a phase perturbation sequence can be constructed, as follows: Step S450: Extract the instantaneous phase information of the dominant frequency components using Hilbert transform, and construct a phase perturbation sequence in chronological order.

[0055] To capture the subtle phase changes of the extracted dominant frequency components over time, an analytic signal is constructed using the Hilbert transform. The Hilbert transform is a linear operator that converts a real signal into a complex analytic signal. It generates the Hilbert transform of the original signal by shifting the positive frequency components of the signal by -90 degrees and the negative frequency components by +90 degrees in the frequency domain.

[0056] The dominant frequency component and its Hilbert transform result are used as the real and imaginary parts, respectively, and combined to form a complex analytic signal. Instantaneous phase information is then extracted from this complex analytic signal. This instantaneous phase information accurately describes the phase angle state of the signal at each sampling moment and can capture minute phase distortions caused by multipath reflections or medium inhomogeneities.

[0057] Finally, the extracted continuous instantaneous phase information is arranged and stored in chronological order to construct a phase perturbation sequence. This phase perturbation sequence, as a high-dimensional time-series data object, not only contains phase values ​​but also implicitly reveals the dynamic perturbation patterns during fault vibration propagation, providing crucial data support for subsequent signal compensation, distortion correction, and fault type identification. Through the above processing, the abstract waveform superposition state is transformed into a quantifiable and calculable phase perturbation sequence, realizing the engineering implementation from signal preprocessing to feature depth extraction.

[0058] Step S150: When the amplitude of the phase perturbation sequence is lower than the preset attenuation threshold, select the corresponding frequency band to perform selective gain compensation on the phase perturbation sequence according to the center frequency response curve corresponding to the dominant frequency component, and obtain the compensated phase perturbation sequence.

[0059] In practical applications, the preset attenuation threshold is a critical threshold for measuring whether the signal strength of the phase perturbation sequence affects the accuracy of subsequent processing. Its setting is based on factors including the background noise level of the sensing system, the statistical characteristics of historical fault data, and the minimum signal-to-noise ratio requirements of the fault identification model.

[0060] In one optional implementation, the preset attenuation threshold is pre-calibrated through the following steps: First, acquire a segment of background noise data from a fault-free distributed fiber optic sensing system and calculate the root mean square value of the noise data as the baseline noise level. Second, perform statistical analysis on the phase disturbance sequences collected from multiple historical grounding fault events, divide the fault samples into different amplitude intervals according to the signal amplitude, calculate the fault type identification accuracy for samples within each amplitude interval, and obtain the signal amplitude-identification accuracy correspondence curve. Then, based on the minimum requirement for identification accuracy of the fault diagnosis system, such as not less than 95%, find the corresponding amplitude critical point in the above signal amplitude-identification accuracy curve. Finally, compare the amplitude critical point with the baseline noise level, take the larger of the two values ​​and multiply it by a preset safety margin coefficient, such as 1.2 to 1.5, as the preset attenuation threshold.

[0061] In another alternative implementation, the preset attenuation threshold can also be theoretically calculated based on the transmission loss model of the optical fiber sensing link: obtain the length of the optical fiber link and the insertion loss parameters of each connection point, calculate the theoretical attenuation of the fault vibration signal propagating along the optical fiber to the sensing end, and use the remaining signal amplitude corresponding to the theoretical attenuation as the preset attenuation threshold.

[0062] The center frequency response curve is a function curve that describes the amplitude response characteristics of a signal within a preset frequency range, based on the dominant frequency component. It reflects the gain or attenuation characteristics of the sensing system or transmission medium to different frequency components. The process of generating this curve is, for example, as follows: using the dominant frequency component as a reference point, a nonlinear fit is performed on the phase perturbation sequence within the preset frequency range to generate the center frequency response curve, thereby accurately locating the frequency bands where signal attenuation is severe.

[0063] Frequency-selective gain compensation enhances signals within specific attenuation frequency bands, avoiding excessive noise amplification caused by uniform amplification across the entire frequency band. In practice, when the amplitude of the phase perturbation sequence is determined to be below a preset attenuation threshold, the frequency band with severe signal attenuation is first identified based on the center frequency response curve, and then selective gain compensation for that specific frequency band is initiated. This compensation process adaptively adjusts the gain coefficient to enhance the signal and correct phase distortion within the attenuation frequency band, ultimately yielding the compensated phase perturbation sequence.

[0064] For details on the implementation of selective gain compensation, please refer to [link / reference]. Figure 5 The diagram shows a flowchart illustrating a method for selectively gain-compensating a phase perturbation sequence by selecting a corresponding frequency band to obtain a compensated phase perturbation sequence, according to an embodiment of this application. The method specifically includes: Step S510: Using the dominant frequency component as a reference point, perform nonlinear fitting on the phase perturbation sequence within a preset frequency range to generate a center frequency response curve.

[0065] Extract the dominant frequency component determined in the previous steps and use it as the central anchor point for constructing the frequency response model. Based on this benchmark, set a preset frequency range covering the dominant frequency component, for example, extending the frequency range upwards and downwards by 10% to 20% centered on the dominant frequency component, to ensure that the fitting process can fully capture the main frequency band features where the signal energy is concentrated.

[0066] Within a defined preset frequency range, the phase values ​​corresponding to each frequency point within this range are extracted from the phase perturbation sequence, forming a discrete frequency-phase data point set, which serves as the input data for nonlinear fitting. Subsequently, a nonlinear fitting operation is performed on the discrete data point set. This fitting process aims to describe the nonlinear dynamic relationship between phase and frequency, rather than a simple linear regression. Nonlinear models such as the Lorentz function or Gaussian function are used to approximate the discrete frequency response data, generating a continuous curve reflecting the phase variation pattern of the current fiber optic sensing system near the dominant frequency, i.e., the center frequency response curve. This center frequency response curve characterizes the inherent frequency response features of the system, and its fitting accuracy directly determines the accuracy of subsequent identification of signal attenuation frequency bands.

[0067] Step S530: Identify the signal attenuation frequency band based on the center frequency response curve and activate the frequency band selective compensation mechanism.

[0068] The generated center frequency response curve is scanned for spectral characteristics to locate regions on the curve where amplitude drops or phase distortion exceed preset levels. These regions typically correspond to signal attenuation frequency bands in fiber optic transmission links caused by connector loss, medium inhomogeneity, or multipath effects. Specifically, the gain slope corresponding to each frequency point on the curve is calculated, and continuous frequency bands with slopes below a preset negative slope threshold and amplitudes below a preset reference level are marked as signal attenuation frequency bands.

[0069] The preset negative slope threshold and preset reference level can be predetermined based on the nominal frequency response characteristics of the fiber optic sensing system. Specifically, during initial system deployment or periodic calibration, a standard test signal with known frequency and amplitude is injected into the sensing fiber to obtain the standard gain response curve of the system at each frequency point; the gain slope value at each frequency point in the standard gain response curve is subtracted from the preset error tolerance and used as the negative slope threshold at that frequency point; the amplitude at each frequency point in the standard gain response curve is multiplied by a preset scaling factor (e.g., 0.7 to 0.9) and used as the reference level at that frequency point.

[0070] In actual operation, the gain slope of the center frequency response curve is compared with the negative slope threshold of the corresponding frequency point, and the amplitude is compared with the reference level of the corresponding frequency point. When both conditions are met, the frequency point is determined to belong to the attenuation band.

[0071] Upon identifying the signal attenuation frequency band, a band-selective compensation mechanism is immediately activated. This mechanism differs from traditional full-band uniform gain amplification; compensation is applied only within the confirmed signal attenuation frequency band. By multiplying the compensation gain coefficient by a frequency domain weighting function, the gain coefficient of the non-attenuated frequency band approaches 1, meaning neither amplification nor attenuation occurs. The gain coefficient of the attenuated frequency band takes a value greater than 1, thus strictly limiting the compensation operation to the frequency band where attenuation occurs, avoiding interference with the normal signal frequency band. This targeted activation strategy ensures that subsequent gain adjustments only apply to the damaged spectral components that need repair, thereby maintaining the natural properties of the overall signal spectral structure.

[0072] Step S550: Within the signal attenuation frequency band, the minimum mean square error algorithm is introduced to perform adaptive adjustment of the gain coefficient, and complex multiplication is performed on the phase perturbation sequence to complete the phase distortion correction and obtain the compensated phase perturbation sequence.

[0073] Based on the active band selective compensation mechanism, a minimum mean square error algorithm is introduced as the core control logic to perform adaptive adjustment of the gain coefficients. This algorithm approximates the optimal filter coefficients by iteratively updating the weight vector based on the error between the input signal and the desired reference signal. In this scenario, the optimization objective is set as minimizing the mean square error between the compensated signal phase and the ideal lossless phase.

[0074] The adaptive adjustment of the gain coefficient constitutes a dynamic feedback closed loop. An initial iteration step size is set, and the current gain coefficient is adjusted in real time according to changes in the error gradient. The value of the initial iteration step size is determined based on the power level of the input signal and the system's comprehensive requirements for convergence speed and steady-state accuracy. A specific method for determining this step size is as follows: calculate the average power of the phase perturbation sequence within the signal attenuation frequency band; set the initial iteration step size to the reciprocal of the average power multiplied by a preset convergence coefficient, where the convergence coefficient ranges from 0.01 to 0.1. This process can be formally expressed as: (1) In the formula, Indicates the first The gain coefficient vector of the next iteration. The step size factor represents the step size that controls the stability and convergence speed of the algorithm. Indicates the first Error signal of the next iteration This represents the conjugate of the input signal. Through this iterative formula, the algorithm gradually converges to the optimal gain coefficient that minimizes the mean square error, thereby avoiding signal clipping due to excessive gain or insufficient compensation due to insufficient gain.

[0075] After determining the optimal gain coefficient, a complex multiplication operation is performed on the phase perturbation sequence within the signal attenuation frequency band. Since the phase perturbation sequence is represented in the frequency or analytic domain as a complex number containing amplitude and phase information, direct complex multiplication with the adaptively adjusted complex gain coefficient simultaneously amplifies the signal amplitude and corrects the phase rotation. This operation accurately cancels out the phase distortion introduced by channel attenuation, restoring the true waveform characteristics of the signal, thus obtaining a fully repaired and compensated phase perturbation sequence. This provides a high-fidelity data foundation for subsequent fault type determination and location calculations.

[0076] By employing frequency band identification based on the center frequency response curve, gain adaptation driven by minimum mean square error, and phase correction in the complex domain, the phase distortion problem of signals in multipath environments is effectively solved, significantly improving the signal processing accuracy of the cable fault location system.

[0077] In one alternative embodiment, such as Figure 6 As shown, this application also provides a further optimized embodiment based on the method of the above embodiments, the method further comprising: Step S610: During the adaptive adjustment of the gain coefficient, monitor the nonlinear distortion index in real time.

[0078] Within the iterative cycle of adaptive gain coefficient adjustment, the nonlinear distortion index is monitored in real time. This nonlinear distortion index quantifies the degree of waveform distortion caused by excessive amplification of the signal. Essentially, it is a high-order statistic characterizing the signal's deviation from linear transfer characteristics, rather than a simple signal quality assessment score. Specifically, the nonlinear distortion index covers parameters such as total harmonic distortion or intermodulation distortion, accurately reflecting the harmonic component energy introduced by the amplifier circuit or digital gain processing stage.

[0079] The process of obtaining this indicator includes: extracting the energy values ​​of the fundamental component and each harmonic component from the frequency domain data obtained in the preceding frequency domain transformation processing, and calculating the ratio of the fundamental energy to the sum of the harmonic energies. As the gain coefficient gradually increases, when the signal gradually approaches or enters the saturation region of the processing unit, the value of the nonlinear distortion indicator will show a nonlinear and rapid upward trend. By monitoring this indicator in real time, that is, immediately performing frequency domain transformation and statistical analysis on the newly output signal segment after each gain iteration update, early signs of clipping can be captured.

[0080] For example, during the iterative process using the least mean square error algorithm, the total harmonic distortion (THD) of the current output signal is calculated immediately after each weight update and generation of a new gain coefficient. If the calculated THD value jumps to a preset alarm threshold, such as exceeding 1.2%, the signal is determined to be in a state of severe nonlinear distortion.

[0081] Specifically, the aforementioned preset alarm threshold is determined in advance based on the system's requirements for signal fidelity and the tolerance of the subsequent fault feature identification model to signal quality. Specifically, multiple sets of sample signals of known fault types are collected, and the total harmonic distortion (THD) value of each set of samples at different gain levels is calculated. The THD value corresponding to the fault feature identification accuracy dropping to a preset threshold (e.g., 98%) is used as the preset alarm threshold.

[0082] Step S630: When the gain coefficient reaches the saturation limit value according to the nonlinear distortion index, stop the gain increase and record the current gain coefficient as the compensation gain saturation limit value and write it into the storage register to constrain the amplification factor of subsequent signals in the same frequency band.

[0083] Based on the above monitoring results, after each gain iteration update, the system performs frequency domain transformation and statistical analysis on the current output signal to calculate the current nonlinear distortion index value. This index changes dynamically as the gain coefficient gradually increases. When the gain coefficient is low, the signal operates in the linear region, and the nonlinear distortion index value changes slowly; as the gain coefficient gradually increases and the signal gradually approaches the saturation region of the processing unit, the nonlinear distortion index value will show a nonlinear and rapid upward trend. The system compares the nonlinear distortion index value calculated in real time with a preset alarm threshold. When the nonlinear distortion index exceeds the preset alarm threshold, it is determined that the current gain coefficient has reached the saturation limit value, and at this time, the action of stopping gain adjustment is triggered. When the gain coefficient reaches the saturation limit value, gain adjustment is immediately stopped. The saturation limit value serves as a safety threshold in the gain coefficient adjustment process, and its physical meaning corresponds to the maximum linear amplification capability of the signal processing channel without significant clipping distortion.

[0084] The saturation limit value is set based on factors including the power supply voltage margin of the hardware circuit, the quantization bit width of the digital-to-analog converter, and the system's preset signal-to-noise ratio degradation tolerance. One specific setting method is as follows: During system debugging, a standard test signal is input into the signal processing channel, and the gain coefficient is gradually increased while simultaneously monitoring the waveform shape and total harmonic distortion (THD) of the output signal. The gain coefficient value at which the output signal waveform begins to show clipping distortion and the THD exceeds a preset alarm threshold is subtracted from a preset safety margin (e.g., 10%), and this value is used as the saturation limit value.

[0085] For example, if the amplification module used in the system causes the output signal amplitude to start hitting the power rail voltage limit when the gain reaches 2.5, resulting in waveform clipping, then the saturation limit value is set to 2.5. When the real-time calculated gain coefficient value, for example, updated iteratively to 2.5 or a critical value greater than 2.5 via a minimum mean square error algorithm, is detected to reach the preset saturation limit value, the current gain coefficient is immediately locked, preventing the algorithm from further increasing the gain value. This mechanism forcibly interrupts the gain increase process that may lead to irreversible signal distortion, ensuring the waveform integrity of the output signal, preventing the introduction of false high-frequency components due to overcompensation, and thus avoiding these false components interfering with subsequent fault characteristic matching.

[0086] While stopping gain adjustment, the current gain coefficient is recorded as the compensation gain saturation limit value and written to the storage register. The storage register, as a non-volatile storage medium or a specific mapped area in system memory, is used to persistently save the state of key parameters during system operation.

[0087] The write operation involves serializing and encoding the gain coefficient (e.g., 2.5) that currently reaches the saturation limit, and storing it in a pre-allocated address space to form a configuration record containing the gain value, the corresponding center frequency of the frequency band, and a recording timestamp. This recording process aims not only to save the current adjustment result but also to establish a historical parameter archive, enabling the system to remember the critical gain level that causes distortion.

[0088] The aforementioned stored compensation gain saturation limit value is used to constrain the amplification factor of subsequent signals in the same frequency band, reflecting the time-domain multiplexing mechanism of the parameters. Once the compensation gain saturation limit value is stored in the register, if a fault vibration signal belonging to the same frequency band (i.e., with the same center frequency or within the preset frequency tolerance range) is captured again in a subsequent time period, the system will directly use the stored limit value as the upper bound of the gain coefficient for initialization or constraint when invoking the frequency band selective compensation mechanism.

[0089] For example, when processing the next frame of a signal with the same center frequency of 2.4 GHz, regardless of the theoretically optimal gain calculated by the adaptive algorithm, the system forcibly limits it to no more than 2.5 times the recorded value. This constraint mechanism avoids repeated attempts at gain levels already identified as oversaturated under the same physical conditions, thus keeping signal processing within the linear operating range. This self-constraining mechanism based on historical feedback effectively reduces the probability of system oscillations or divergence, ensuring long-term stability and reliability for processing signals with similar characteristics.

[0090] Step S170: Determine the instantaneous fault type based on the compensated phase disturbance sequence, and calculate the propagation distance of the vibration wave in the cable path by combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay, and the arrival timestamp, to obtain the initial positioning coordinates.

[0091] Transient fault types include single-phase grounding faults, arcing grounding faults, and intermittent discharge faults. Identifying the type helps in selecting a matching physical parameter model. Fiber optic medium propagation speed correction parameters are used to calibrate the actual propagation speed of vibration waves in the fiber to compensate for the effects of differences in material refractive index, coating structure, and cabling processes. Total branch node propagation delay refers to the total additional propagation time of vibration waves when passing through nodes such as branch boxes and intermediate joints due to path switching and structural abrupt changes.

[0092] Specifically, such as Figure 7 The diagram shown is a flowchart of a method for determining the instantaneous fault type based on a compensated phase perturbation sequence according to an embodiment of this application. The method includes: Step S710: Perform frequency domain transformation on the compensated phase perturbation sequence to construct a frequency domain energy redistribution feature vector.

[0093] After obtaining the compensated phase perturbation sequence through gain compensation and phase distortion correction, a frequency domain transformation operation is performed on the sequence to extract structured high-dimensional abstract features characterizing the energy distribution of the fault vibration signal in different frequency bands. This frequency domain energy redistribution feature vector can reflect the specific response mode of the mechanical impact caused by the fault in the spectrum. In specific implementation, algorithms such as Fast Fourier Transform or Discrete Cosine Transform can be used to transform the compensated phase perturbation sequence in the time domain to the frequency domain, and the energy value of each sub-band can be calculated according to a preset frequency division strategy.

[0094] For example, the effective frequency band of the compensated phase perturbation sequence (e.g., 0 to 5000 Hz) can be divided into a predetermined number of sub-bands at equal intervals. The sum of squares of the signal amplitude within each sub-band is calculated to obtain the energy value of each sub-band. Subsequently, these energy values ​​are arranged in frequency order to form a frequency domain energy redistribution feature vector. Through the above frequency domain energy statistics method, phase information that may have jitter in time can be transformed into stable energy distribution characteristics, thereby providing a data foundation with strong anti-interference capabilities for subsequent pattern matching.

[0095] Step S730: Calculate the Euclidean distance between the frequency domain energy redistribution feature vector and the energy distribution vectors of each historical fault in the pre-established instantaneous fault feature library.

[0096] The pre-established instantaneous fault feature library is a dataset built during the system's offline training or initialization phase. Its data source is a large number of fault samples accumulated in the historical fault database. This feature library collects and labels vibration signal samples under different types of faults (such as single-phase ground faults, instantaneous arc discharges, etc.), performs the same frequency domain energy redistribution feature extraction operation as the current step on each sample, generates historical fault energy distribution vectors corresponding to each category, and finally forms a database with fault categories as indexes and feature vectors as content.

[0097] Based on the real-time generated frequency domain energy redistribution feature vector, it is compared one by one with the historical fault energy distribution vectors stored in the instantaneous fault feature database, and the Euclidean distance between the two is calculated. Euclidean distance measures the straight-line distance between two multidimensional vectors in geometric space; the smaller the distance value, the higher the similarity between the two vectors, meaning the current vibration signal is closer to the characteristics of that historical fault type. This distance metric quantifies the degree of difference between the current fault state and known fault types.

[0098] Step S750: Convert the Euclidean distance into fault feature matching degree using a linear mapping algorithm.

[0099] A linear mapping algorithm is used to convert the calculated Euclidean distance values ​​into normalized fault feature matching scores for easier intuitive judgment. The fault feature matching score is a dimensionless percentage value reflecting the probability confidence that the current signal belongs to a certain type of fault. The linear mapping algorithm is based on the negative correlation between Euclidean distance and matching score; that is, the smaller the distance, the higher the matching score. This mapping relationship can be expressed as: (2) in, For fault feature matching degree, The calculated Euclidean distance, This is a preset maximum distance reference value. This maximum distance reference value can be set based on the maximum dispersion of historical data, or it can be taken as the diagonal length in the feature space. Through this transformation, abstract distance values ​​are mapped to easily understood percentage indicators, making subsequent threshold determination logic more intuitive and easier to engineer.

[0100] Step S770: When the fault feature matching degree is greater than the preset matching degree threshold, it is determined that an instantaneous fault has occurred, and the timestamp corresponding to the compensated phase disturbance sequence is output as the instantaneous fault occurrence time, and the fault type with the highest fault feature matching degree is output as the instantaneous fault type.

[0101] The preset matching threshold is a decision threshold that distinguishes between normal fluctuations and real instantaneous faults. This threshold is determined by collecting historical test data, plotting precision-recall curves for each fault category, and finding the optimal balance between false negatives and false positives based on the tolerance of the business scenario. For example, for high-risk faults such as single-phase-to-ground short circuits, the threshold can be appropriately lowered to improve recall; while for minor interferences, a higher threshold is set to avoid false positives.

[0102] The system monitors the matching degree of fault features in real time. When the matching degree exceeds a preset threshold, a transient fault is identified in the monitored signal. Subsequently, the system selects the fault label corresponding to the historical fault energy distribution vector with the highest matching degree from the feature library, classifying it as the transient fault type. Simultaneously, the system parses the data packet header or metadata of the compensated phase perturbation sequence, extracts the absolute time corresponding to the signal's generation, records it, and outputs it as the transient fault occurrence time. This determination logic ensures that the output of the fault type has clear statistical basis, thereby improving the reliability of fault identification.

[0103] In the aforementioned technical process, the frequency domain energy redistribution feature vector provides a robust feature foundation, the pre-established instantaneous fault feature library provides a standard reference for identification, Euclidean distance quantifies the similarity between features, the linear mapping algorithm transforms geometric distance into a decisionable confidence index, and the preset matching threshold serves as the final decision switch. These features work synergistically to achieve accurate mapping from phase perturbation sequences to specific fault types. By constructing a complete link of feature extraction, distance measurement, confidence conversion, and threshold determination, the process effectively solves the problems of strong subjectivity and susceptibility to environmental noise interference in manual identification, ensuring accurate and rapid identification of fault types and occurrence times even in complex cable operating environments.

[0104] Furthermore, such as Figure 8 The diagram shown is a flowchart of a method for calculating the propagation distance of a vibration wave in a cable path by combining optical fiber medium propagation speed correction parameters, according to an embodiment of this application. The method includes: Step S810: Obtain the optical fiber medium propagation speed correction parameters and the real-time temperature value of the environment where the sensing optical fiber is located. Substitute the real-time temperature value into the preset temperature-sound speed coupling model to calculate the temperature-affected speed compensation factor.

[0105] The optical fiber propagation speed correction parameter is a reference physical quantity determined based on the refractive index of the fiber core and the elasto-optic coefficient. It is typically set according to factory calibration or laboratory measurements and is used to characterize the propagation speed of vibration waves in optical fibers under standard conditions. Real-time temperature values ​​are acquired by temperature sensing modules deployed along the sensing fiber, such as Raman scattering temperature measurement units or fiber Bragg grating sensors. The data unit is usually degrees Celsius (°C). Using real-time monitored ambient temperature data, the propagation characteristics of vibration waves in the optical fiber medium can be dynamically corrected, eliminating sound speed drift caused by environmental thermal effects.

[0106] The temperature-velocity-of-sound coupling model describes the mapping relationship between ambient temperature and sound wave propagation speed, and its construction is completed during the system calibration phase. By collecting experimental data on sound wave propagation in optical fibers under different temperature gradients (e.g., -10℃ to 60℃), and based on the physical laws governing sound speed with temperature (usually linear or polynomial), the least squares method is used to fit the function curve between temperature and the sound speed correction factor, thus forming the preset temperature-velocity-of-sound coupling model. Substituting the real-time acquired temperature values ​​into this model allows the derivation of the temperature-affected velocity compensation factor under the current environment. This temperature-affected velocity compensation factor is used to fine-tune the reference propagation speed to obtain the true propagation speed under the current operating conditions.

[0107] Step S830: Determine the actual propagation speed of the vibration wave in the optical fiber medium based on the optical fiber medium propagation speed correction parameter and the temperature influence speed compensation factor.

[0108] The temperature-affected velocity compensation factor is fused with the fiber optic medium propagation velocity correction parameter to determine the actual propagation velocity of the vibration wave in the fiber optic medium under the current specific temperature environment. This actual propagation velocity is the core kinematic parameter for subsequent fault distance calculation. By introducing the temperature-affected velocity compensation factor, the fiber elasto-optic effect and refractive index changes caused by ambient temperature variations can be effectively corrected, avoiding the accumulation of positioning errors caused by velocity reference drift. Specifically, the determination process typically involves multiplying the fiber optic medium propagation velocity correction parameter by the temperature-affected velocity compensation factor, or performing a weighted summation of the two according to a preset coupling algorithm, thereby outputting the dynamically corrected actual propagation velocity.

[0109] Step S850: Obtain the number of branch nodes and the corresponding single-node propagation delay in the branch cable topology, and calculate the total branch node propagation delay.

[0110] The branch cable topology reflects the connection relationships and physical layout of the nodes in the cable network. The number of branch nodes can refer to the total number of cable branch boxes, junction boxes, or other connection points passed through on the vibration wave propagation path. The propagation delay per node can refer to the fixed time loss caused by structural discontinuities, medium interface reflections, or scattering when the vibration wave passes through each branch node. This parameter is usually pre-calibrated and stored based on historical test data or the physical characteristics of the node type.

[0111] To quantify the additional delay effect of cable network topology on vibration wave propagation time, it is necessary to retrieve the list of branch nodes included in the current fault path from a pre-configured cable topology database, extract the single-node propagation delay corresponding to each node in the list, and sum them up to obtain the total branch node propagation delay. This total delay characterizes the non-propagational time lag caused solely by node structure factors when the vibration wave traverses the entire cable network topology, and must be eliminated during distance calculation.

[0112] Step S870: Calculate the difference between the instantaneous fault occurrence time and the arrival timestamp to obtain the total propagation time of the vibration wave, and subtract the total branch node propagation delay from the total propagation time to obtain the effective propagation time.

[0113] The instantaneous fault occurrence time is the absolute time at which the vibration signal is generated at the fault point determined in the preceding steps, while the arrival timestamp is the moment when the vibration signal is captured by the distributed fiber optic sensing device. The time difference between the two is the total propagation time, which characterizes the overall duration from the fault occurrence to the signal detection. To extract the propagation time purely for the spatial travel of the vibration wave, the structural delay caused by the topological nodes needs to be removed from the total propagation time.

[0114] Specifically, the total propagation time is corrected using the total branch node propagation delay calculated in step S850, i.e., the total branch node propagation delay is subtracted from the total propagation time. This process eliminates non-propagation time interference introduced by the nodes, thereby obtaining the effective propagation time of the vibration wave in the optical fiber medium. This effective propagation time directly corresponds to the flight time of the vibration wave on the cable path, forming an accurate time reference for calculating the distance.

[0115] Step S890: Multiply the actual propagation speed by the effective propagation time and combine it with the cable path bending coefficient to obtain the propagation distance of the vibration wave in the cable path.

[0116] The product of actual propagation speed and effective propagation time theoretically corresponds to the equivalent straight-line distance of vibration wave propagation. However, in actual cable laying, cables often exhibit complex geometric shapes such as bends and detours, and do not extend in a straight line. To reflect the actual physical laying path, a cable path bending coefficient is introduced for correction.

[0117] The cable path bending coefficient is a dimensionless parameter characterizing the proportional relationship between the actual cable laying length and the equivalent straight length. Its value is greater than or equal to 1, and it is calculated statistically based on geographic information system data, design drawings, or measured path lengths of the cable path. Multiplying the intermediate result obtained by multiplying the actual propagation speed by the effective propagation time, and then multiplying it by the cable path bending coefficient, maps the straight-line propagation distance to the actual propagation distance considering the bending path. This actual propagation distance accurately characterizes the physical distance from the fault point along the actual cable laying path to the monitoring end, providing high-precision distance input for subsequent coordinate calculations.

[0118] By combining dynamic sound velocity correction based on real-time temperature data, precise stripping of branch node delays, and geometric mapping of cable path bending coefficients, these steps work synergistically to minimize the impact of environmental thermal noise interference, topology errors, and path geometric distortions on the location results. The temperature-sound velocity coupling model and node delay eliminate non-propagational errors in time measurement from the perspectives of medium properties and network structure, respectively, while the cable path bending coefficient completes the final conversion from theoretical straight-line distance to actual physical distance. This multi-dimensional joint correction mechanism ensures that the calculated vibration wave propagation distance accurately reflects the spatial location of the fault point in the complex cable network, laying a reliable data foundation for generating high-precision initial location coordinates, significantly reducing location errors in complex cable network environments, and achieving fault location accuracy at the meter or even sub-meter level.

[0119] Furthermore, after obtaining the propagation distance, using the known starting point of the cable line, such as the installation location of a substation or cable branch box, as the spatial reference origin, and combining it with the cable's vector direction model—that is, a digital description of the cable laying path, including the direction and length information of each segment—the propagation distance is extrapolated along the cable's direction to determine the location of the fault point in space. Specifically, starting from the reference origin, the distance is accumulated segment by segment along the cable's direction vector until the accumulated distance equals the calculated propagation distance. The spatial location reached at this point is the initial location coordinate of the fault point. This coordinate is output in latitude and longitude form or in planar coordinate form relative to the reference point.

[0120] In this way, the one-dimensional propagation distance is converted into positioning coordinates with practical spatial significance, providing an operable positional reference for subsequent error assessment and correction.

[0121] Step S190: If the error of the evaluated initial positioning coordinates exceeds the preset boundary, the anchor point is queried according to the instantaneous fault type, and the cable path is interpolated and corrected using the multimodal wave velocity difference coefficient, and the final positioning coordinates are output.

[0122] The preset boundary is the maximum spatial range of allowable positioning error, which is set based on the accuracy requirements of power distribution network emergency repair, the coordinate resolution of the GIS (Geographic Information System), and the statistical distribution of historical positioning data errors.

[0123] Specifically, such as Figure 9 As shown in the embodiments of this application, a specific method for evaluating whether the error of the initial positioning coordinates exceeds a preset boundary is also provided. This method includes: Step S910: Calculate the covariance matrix of the initial positioning coordinates, and determine the current positioning error based on the covariance matrix.

[0124] In the cable branch box grounding fault location system, although the initial location coordinates calculated in the aforementioned steps macroscopically indicate the geographical location of the fault, the coordinate values ​​are not absolutely certain values ​​due to random factors such as sensor acquisition noise, nonlinear changes in fiber refractive index, and ambient temperature disturbances. Instead, they are estimated values ​​of random variables that follow a specific probability distribution.

[0125] To quantify this uncertainty, a covariance matrix corresponding to the initial positioning coordinates is constructed or retrieved. This covariance matrix describes the dimensions of the multidimensional random variable, typically including the degree of covariance between longitude, latitude, and depth. The diagonal elements represent the positioning variance in the longitude, latitude, and depth directions, respectively, while the off-diagonal elements reflect the correlation between different coordinate dimensions, such as the positive correlation between longitude and latitude errors caused by the cable laying path.

[0126] Based on the covariance matrix, its trace, i.e., the sum of its diagonal elements, is extracted, and the square root of the trace is calculated as a comprehensive positioning error index; alternatively, the direction of the eigenvector corresponding to the largest eigenvalue of the covariance matrix is ​​selected as the direction of the maximum error, and its square root is determined as the current positioning error. Through the above processing, the uncertainty of the positioning result is transformed into a measurable value, providing a statistical basis for subsequent accuracy assessment.

[0127] Step S930: Compare the current positioning error with the preset spatial positioning accuracy error boundary threshold.

[0128] After obtaining the quantified current positioning error, it is compared with a preset spatial positioning accuracy error boundary threshold. This preset spatial positioning accuracy error boundary threshold serves as a decision threshold for determining whether the positioning result meets the requirements of engineering applications. Its setting is based on the operation and maintenance strategy of the power distribution system, the emergency repair resource response capability, and the geographical environment of the cable trench.

[0129] For example, in densely populated commercial areas or core areas with extremely high requirements for power supply reliability, the threshold can be set to a lower value (e.g., 1.5 meters) to ensure that repair personnel can quickly and accurately locate the point. In areas with open terrain, clear cable routes, and abundant resources, the threshold can be appropriately relaxed. The threshold can be determined based on the statistical distribution of errors in successful fault location data from historical fault data, taking the upper limit of the 95% confidence interval as the boundary value; alternatively, it can be derived by combining voltage level and the physical dimensions of the branch box, for example, setting it based on half the length and width of the branch box. By comparing the current positioning error with the threshold, if the current positioning error is within the threshold range, it indicates that the positioning result has sufficient confidence; if it exceeds the threshold, it indicates that the positioning result is greatly disturbed and there is a risk of significant deviation from the true location.

[0130] Step S950: If the current positioning error is greater than the preset spatial positioning accuracy error boundary threshold, then it is determined that the error of the initial positioning coordinates exceeds the preset boundary.

[0131] When the current positioning error exceeds a preset spatial positioning accuracy error boundary threshold, the error of the initial positioning coordinates is determined to exceed the preset boundary. This determination logic transforms continuous error measurement into a clear correction trigger signal, realizing a binary decision on the reliability of the positioning result.

[0132] For example, if the calculated current positioning error is 2.80 meters, while the preset spatial positioning accuracy error boundary threshold is 1.50 meters, the system determines that the current initial positioning coordinates do not meet the requirements for direct release because 2.80 meters is greater than 1.50 meters. This determination serves as a critical status indicator, directly controlling the subsequent algorithm process: if the error does not exceed the boundary, the initial positioning coordinates are directly output as the final result; if the error exceeds the boundary, subsequent steps, including multimodal wave velocity difference coefficient correction and interpolation anchor point query, are triggered to finely calibrate the initial coordinates. This determination mechanism based on statistical error boundaries can adaptively identify and handle scenarios with insufficient positioning accuracy, avoiding low-precision data from misleading maintenance decisions, thereby improving the overall robustness and engineering practical value of the cable fault location system.

[0133] In this embodiment, the correction anchor point is a reference point with precise geographic coordinates used to calibrate the initial positioning results. The instantaneous fault type is used as an index to query the correction anchor point because different fault types exhibit different energy attenuation and waveform distortion characteristics along their propagation paths, corresponding to different optimal anchor point combinations. The multimodal wave velocity difference coefficient reflects the difference in propagation speed of different frequency components in the vibration signal within the same medium, and is used to correct the deviation between the straight-line distance and the actual arc length caused by the dispersion effect of different frequency signals on the cable path. Interpolation correction is a process of finely adjusting the initial positioning coordinates using the known spatial constraint relationships between anchor points.

[0134] Specifically, such as Figure 10 The diagram shows a flowchart of a method provided in this application for querying and correcting anchor points based on instantaneous fault types, and interpolating and correcting cable paths using multimodal wave velocity difference coefficients to output final positioning coordinates. The method includes: Step S1010: Based on the instantaneous fault type, match and extract the coordinates of the two known anchor points closest to the initial positioning coordinates from the preset anchor point database, and use them as correction anchor points.

[0135] The pre-built anchor point database is a data set pre-constructed and stored in the system's storage medium, containing spatial location information of known feature points along the cable path. The database formation process includes: during system initialization or cable laying acceptance phases, using high-precision measuring equipment to collect the latitude and longitude coordinates and depth information of key nodes along the cable route, such as corner wells, joint wells, and branch box entrances; subsequently, organizing this data according to the cable topology to form a structured data table with node identifiers as the primary key and coordinate components as fields.

[0136] When it is necessary to correct the initial positioning coordinates, the instantaneous fault type determined in the previous steps is first read, such as single-phase ground fault or instantaneous flashover, and this fault type is used as a search condition to filter historical feature point data or full-line reference anchor point data related to this type from the preset anchor point database. Based on this, a spatial neighborhood search algorithm is used to calculate the Euclidean distance between the initial positioning coordinates and the coordinates of each candidate anchor point in the database. The two anchor points with the smallest distance values ​​are selected and defined as the forward correction anchor point and the backward correction anchor point, collectively referred to as the correction anchor points. By limiting the interpolation correction to an interval with similar local physical characteristics, the additional errors introduced by changes in cable medium parameters during interpolation over long distances can be effectively avoided.

[0137] Step S1030: Extract the high-frequency and low-frequency components from the traveling wave signal, calculate the difference in propagation speed between the two, and obtain the multimode wave speed difference coefficient.

[0138] Traveling wave signals can refer to the time-domain waveform data sensed in distributed optical fiber sensing devices when mechanical vibration waves generated at a fault point propagate along the cable medium. Multi-scale decomposition processing is performed on the acquired raw traveling wave signals, such as using wavelet transform or empirical mode decomposition, to decompose the signal into sub-signals representing different frequency components.

[0139] Among them, the high-frequency component corresponds to the short-wavelength, rapidly changing details in the signal; the low-frequency component corresponds to the longer-wavelength, gently changing main body. After extracting the above components, the propagation characteristics of the high-frequency and low-frequency components in the cable medium are tracked separately. By calculating the time difference or phase difference between the two reaching the same monitoring position, and combining it with the propagation distance, the high-frequency wave velocity and low-frequency wave velocity are calculated respectively.

[0140] The multimodal wave velocity difference coefficient is a quantitative index calculated based on the degree of difference between high-frequency and low-frequency wave velocities. Its calculation logic can be either the ratio of the difference between the two to the average wave velocity, or the calculation of their dispersion factors. This coefficient reflects the dispersion effect of the cable medium on different frequency components; that is, the non-uniformity of propagation speeds caused by the different response characteristics of the medium for different frequency components. This dispersion characteristic is the physical root cause of systematic errors in positioning based on a single wave velocity. Introducing the multimodal wave velocity difference coefficient can transform the physical waveform dispersion difference into a geometric path correction parameter.

[0141] Step S1050: Obtain the actual laying curvature radius of the cable in the geographic information system, and use the cubic spline interpolation algorithm to substitute the multimodal wave velocity difference coefficient into the arc length integral formula between the correction anchor points to calculate the bending length correction value of the cable path.

[0142] A GIS system is a computer system used to collect, store, manage, analyze, and display geospatial data. This application primarily provides digital geographic information about cable laying paths. The actual cable laying curvature radius can refer to a measure of the degree of physical bending of the cable in underground trenches, tunnels, or direct burial scenarios due to terrain limitations or construction techniques. This data can be directly read from the cable vector layer in the GIS system or calculated through node coordinate fitting. Cubic spline interpolation is a piecewise polynomial interpolation method that ensures the continuity of function values, first derivatives, and even second derivatives at each correction anchor point, thereby fitting a smooth cable path curve that conforms to physical constraints.

[0143] A parameterized curve equation connecting two corrected anchor points is constructed using a cubic spline interpolation algorithm. The independent variable of this equation is the arc length parameter along the path. Subsequently, the multimodal wave velocity difference coefficient obtained in the preceding steps is used as a dynamic weighting factor or correction operator and substituted into the arc length integral formula established based on the aforementioned curve equation. This arc length integral formula describes the nonlinear mapping relationship between the theoretical straight-line distance and the actual curved propagation distance, considering the multimodal wave velocity dispersion effect. By numerically solving this integral formula, the deviation between the theoretical and actual distances is calculated, thus obtaining the cable path length correction value. This process fully utilizes the spatial geometric information provided by GIS and the physical dispersion information extracted by signal processing, achieving a joint correction based on both geometry and physics.

[0144] Step S1070: The bending length correction value of the cable path is used as a compensation vector and superimposed on the tangent direction of the initial positioning coordinates. The final positioning coordinates are obtained through coordinate transformation.

[0145] The tangential direction can refer to the unit vector extending along the theoretical path of the cable at the initial positioning coordinate point. This direction represents the geometric orientation of the fault point relative to the monitoring system. The cable path bending length correction value, as a scalar parameter, first needs to be converted into a compensation vector with directional attributes. This is achieved by multiplying the correction value by the tangential direction unit vector to generate a spatial correction. After obtaining this compensation vector, it is superimposed onto the initial positioning coordinates to achieve the spatial displacement of the coordinate point.

[0146] Considering the potential differences in definition between the initial positioning coordinate system (e.g., a one-dimensional coordinate system based on fiber length) and the final output coordinate system (e.g., a geographic latitude and longitude coordinate system or a local construction coordinate system), the system uses a preset coordinate transformation matrix or transformation algorithm to convert the superimposed and corrected intermediate coordinates into values ​​in the target coordinate system. The final positioning coordinates, generated after the aforementioned geometric offset and coordinate transformation processing, accurately reflect the true geographical location of the fault point. These coordinates effectively eliminate positioning deviations caused by cable bending, multimodal dispersion, and the accumulation of initial calculation errors, making the positioning results approximate the physical reality from the theoretical calculation values.

[0147] Through the above steps, the method provided in this application selects matching anchor point benchmarks based on fault type, constructs a multimode wave velocity difference coefficient using the propagation speed difference between high-frequency and low-frequency traveling wave components, combines the actual cable laying curvature radius in the geographic information system, and accurately fits the cable path using a cubic spline interpolation algorithm. It also incorporates physical dispersion characteristics into the arc length integral calculation, thereby generating an accurate cable path bending length correction value. This correction value, in the form of a compensation vector, spatially calibrates the initial positioning coordinates along the tangential direction, and outputs the final positioning coordinates after coordinate transformation. This effectively solves the positioning error problem caused by the cable path and achieves sub-meter level accurate positioning of grounding faults in complex laying environments.

[0148] Furthermore, an exemplary embodiment of this application also provides a rapid location system for grounding faults in cable branch boxes.

[0149] Figure 11 A block diagram illustrating an exemplary embodiment of a cable branch box grounding fault rapid location system according to this application is shown. (Reference) Figure 11 The cable branch box grounding fault rapid location system 1100 according to an exemplary embodiment of this application may include: a signal acquisition and interference determination module 1110, a frequency domain separation and phase extraction module 1130, a selective gain compensation module 1150, an initial location module 1170, and a path interpolation correction module 1190.

[0150] Specifically, the signal acquisition and interference determination module 1110 can be used to acquire fault vibration signals captured by distributed optical fiber sensing devices deployed along the branch cable, extract the arrival timestamp of the fault vibration signals, and determine whether the fault vibration signals are in a waveform superposition state caused by multipath reflection interference; the frequency domain separation and phase extraction module 1130 can be used to perform frequency domain separation on the fault vibration signals if waveform superposition exists, extract the dominant frequency components, and construct a phase disturbance sequence; the selective gain compensation module 1150 can be used to, when the amplitude of the phase disturbance sequence is lower than a preset attenuation threshold, calculate the phase disturbance response curve based on the center frequency response curve corresponding to the dominant frequency component. The system selects the corresponding frequency band to selectively compensate the phase disturbance sequence, obtaining a compensated phase disturbance sequence. The initial positioning module 1170 can determine the instantaneous fault type based on the compensated phase disturbance sequence, and calculate the propagation distance of the vibration wave in the cable path by combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay, and the arrival timestamp, to obtain the initial positioning coordinates. The path interpolation correction module 1190 can query the correction anchor point according to the instantaneous fault type if the error of the evaluated initial positioning coordinates exceeds the preset boundary, and use the multimode wave velocity difference coefficient to interpolate and correct the cable path, outputting the final positioning coordinates.

[0151] Since the functional modules of the cable branch box grounding fault rapid location system in this application are the same as those in the above method implementation, they will not be described again here.

[0152] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A method for rapid location of grounding faults in cable branch boxes, characterized in that, The method includes: The fault vibration signal captured by the distributed optical fiber sensing device deployed along the branch cable is obtained, the arrival timestamp of the fault vibration signal is extracted, and it is determined whether the fault vibration signal has a waveform superposition state caused by multipath reflection interference. If the waveform superposition state exists, the fault vibration signal is separated in the frequency domain, the dominant frequency component is extracted, and a phase perturbation sequence is constructed. When the amplitude of the phase perturbation sequence is lower than the preset attenuation threshold, the corresponding frequency band is selected according to the center frequency response curve corresponding to the dominant frequency component to perform selective gain compensation on the phase perturbation sequence, thereby obtaining the compensated phase perturbation sequence. The instantaneous fault type is determined based on the compensated phase disturbance sequence, and the propagation distance of the vibration wave in the cable path is calculated by combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay, and the arrival timestamp to obtain the initial positioning coordinates. If the error of the assessed initial positioning coordinates exceeds the preset boundary, the anchor point is queried and corrected according to the instantaneous fault type, and the cable path is interpolated and corrected using the multimodal wave velocity difference coefficient to output the final positioning coordinates.

2. The method according to claim 1, characterized in that, The determination of whether the fault vibration signal has a waveform superposition state caused by multipath reflection interference includes: After background noise filtering of the fault vibration signal, the envelope of the processed signal in the time domain is extracted, and the time interval between adjacent peaks on the envelope is calculated. When a preset number of consecutive time intervals are all less than a preset time threshold, and the amplitude of the secondary peak reaches a preset ratio to the amplitude of the primary peak, it is determined that the waveform superposition state exists.

3. The method according to claim 2, characterized in that, The step of performing frequency domain separation on the fault vibration signal, extracting the dominant frequency components, and constructing a phase perturbation sequence includes: After applying a window function to the processed signal, a fast Fourier transform is performed to obtain the frequency domain energy distribution; Traverse the frequency domain energy distribution, identify frequency segments whose energy proportion exceeds a preset proportion threshold as dominant frequency segments, and extract the corresponding dominant frequency components; The instantaneous phase information of the dominant frequency component is extracted using Hilbert transform, and the phase perturbation sequence is constructed in chronological order.

4. The method according to claim 1, characterized in that, The step of selectively gain compensating the phase perturbation sequence by selecting a corresponding frequency band based on the center frequency response curve corresponding to the dominant frequency component, to obtain a compensated phase perturbation sequence, includes: Using the dominant frequency component as a reference point, the phase perturbation sequence is nonlinearly fitted within a preset frequency range to generate a center frequency response curve. Identify the signal attenuation frequency band based on the center frequency response curve and activate the frequency band selective compensation mechanism. Within the signal attenuation frequency band, a minimum mean square error algorithm is introduced to perform adaptive adjustment of the gain coefficient, and complex multiplication is performed on the phase perturbation sequence to complete phase distortion correction, thereby obtaining the compensated phase perturbation sequence.

5. The method according to claim 1, characterized in that, The step of determining the instantaneous fault type based on the compensated phase perturbation sequence includes: The compensated phase perturbation sequence is transformed in the frequency domain to construct a frequency domain energy redistribution feature vector; Calculate the Euclidean distance between the frequency domain energy redistribution feature vector and each historical fault energy distribution vector in the pre-established instantaneous fault feature database; The Euclidean distance is converted into a fault feature matching degree using a linear mapping algorithm; When the fault feature matching degree is greater than the preset matching degree threshold, an instantaneous fault is determined to have occurred, and the timestamp corresponding to the compensated phase perturbation sequence is output as the instantaneous fault occurrence time. The fault type with the highest fault feature matching degree is output as the instantaneous fault type.

6. The method according to claim 5, characterized in that, The calculation of the propagation distance of the vibration wave in the cable path, combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay, and the arrival timestamp, includes: The optical fiber medium propagation speed correction parameter and the real-time temperature value of the environment where the sensing optical fiber is located are obtained. The real-time temperature value is substituted into the preset temperature-sound speed coupling model to calculate the temperature influence speed compensation factor. The actual propagation speed of the vibration wave in the optical fiber medium is determined based on the optical fiber medium propagation speed correction parameter and the temperature influence speed compensation factor. The number of branch nodes and the corresponding single-node propagation delay in the branch cable topology are obtained, and the total branch node propagation delay is calculated. The difference between the instantaneous fault occurrence time and the arrival timestamp is calculated to obtain the total propagation time of the vibration wave, and the total branch node propagation delay is subtracted from the total propagation time to obtain the effective propagation time; Multiply the actual propagation speed by the effective propagation time, and combine this with the cable path bending coefficient to obtain the propagation distance of the vibration wave in the cable path.

7. The method according to claim 1, characterized in that, Evaluating whether the error of the initial positioning coordinates exceeds a preset boundary includes: Calculate the covariance matrix of the initial positioning coordinates, and determine the current positioning error based on the covariance matrix; The current positioning error is compared with a preset spatial positioning accuracy error boundary threshold. If the current positioning error is greater than the preset spatial positioning accuracy error boundary threshold, then it is determined that the error of the initial positioning coordinates exceeds the preset boundary.

8. The method according to claim 1, characterized in that, The process of querying and correcting anchor points based on the instantaneous fault type, interpolating and correcting the cable path using a multimodal wave velocity difference coefficient, and outputting the final positioning coordinates includes: Based on the instantaneous fault type, the coordinates of the two known anchor points closest to the initial positioning coordinates are matched and extracted from the preset anchor point database and used as the correction anchor points. Extract the high-frequency and low-frequency components from the traveling wave signal, calculate the difference in their propagation speeds, and obtain the multimode wave speed difference coefficient. The actual curvature radius of the cable in the geographic information system is obtained, and the multimodal wave velocity difference coefficient is substituted into the arc length integral formula between the correction anchor points using a cubic spline interpolation algorithm to calculate the correction value of the cable path bending length. The bending length correction value of the cable path is used as a compensation vector and superimposed on the tangent direction of the initial positioning coordinates. The final positioning coordinates are then obtained through coordinate transformation.

9. The method according to any one of claims 1 to 8, characterized in that, The step of acquiring fault vibration signals captured by distributed fiber optic sensing devices deployed along the branch cable and extracting the arrival timestamps of the fault vibration signals includes: The distributed optical fiber sensing device collects continuous backscattered light data streams along the branch cable in real time at a preset sampling rate. A short time window and a long time window are set to perform real-time analysis on the continuous backscattering data stream and calculate the energy ratio of the short time window to the long time window; When the energy ratio is greater than or equal to a preset trigger threshold, it is determined that the fault vibration signal generated at the moment of the fault is captured, and the absolute time corresponding to the sudden change point of the fault vibration signal is extracted as the arrival timestamp.

10. A rapid location system for grounding faults in cable branch boxes, characterized in that, The system includes: The signal acquisition and interference determination module is used to acquire the fault vibration signal captured by the distributed optical fiber sensing device deployed along the branch cable, extract the arrival timestamp of the fault vibration signal, and determine whether the fault vibration signal has a waveform superposition state caused by multipath reflection interference. The frequency domain separation and phase extraction module is used to perform frequency domain separation on the fault vibration signal, extract the dominant frequency components, and construct a phase perturbation sequence if the waveform superposition state exists. The selective gain compensation module is used to select a corresponding frequency band to selectively compensate the phase perturbation sequence when the amplitude of the phase perturbation sequence is lower than a preset attenuation threshold, based on the center frequency response curve corresponding to the dominant frequency component, to obtain a compensated phase perturbation sequence. The initial positioning module is used to determine the instantaneous fault type based on the compensated phase disturbance sequence, and calculate the propagation distance of the vibration wave in the cable path by combining the optical fiber medium propagation speed correction parameter, the total branch node propagation delay and the arrival timestamp, so as to obtain the initial positioning coordinates. The path interpolation correction module is used to query the correction anchor point according to the instantaneous fault type if the error of the evaluated initial positioning coordinates exceeds the preset boundary, and to interpolate and correct the cable path using the multimodal wave velocity difference coefficient, and output the final positioning coordinates.

Images