Traveling wave positioning method, system and equipment based on intelligent fuse, and medium

Abstract

The invention relates to the technical field of traveling wave positioning, and discloses a traveling wave positioning method, system and device based on an intelligent fuse, and a medium. The method comprises the following steps: installing an intelligent fuse high-frequency sampling signal to obtain digital data; the data is transmitted to a master control fuse to synthesize a zero sequence quantity to obtain a fault judgment result; constructing a feature matrix by applying adaptive threshold wavelet transform based on the zero sequence quantity; combining zero-sequence current and traveling wave characteristics and adopting an improved algorithm to carry out distance measurement and fault positioning; and executing hierarchical protection according to a judgment result and a positioning result to obtain maintenance information. According to the intelligent device, the position of a fault point can be accurately positioned, and differential protection actions can be automatically executed according to the severity of the fault, so that the fault searching and processing time is greatly shortened, and the operation reliability and the maintenance efficiency of the power distribution network are improved.

Description

Technical Field

[0001] The present application relates to the technical field of traveling wave positioning, and particularly to a traveling wave positioning method, system, device and medium based on an intelligent fuse. Background Art

[0002] In the distribution network of the power system, traditional fuses are mainly used for overload protection and short-circuit protection, and cut off the fault current by means of fuse melting to avoid safety accidents such as equipment damage and fire. Most of these traditional fuses adopt simple mechanical structures and rely on the thermal effect or electrodynamic effect of current to achieve the protection function. The types include tubular fuses, knife-shaped fuses and drop fuses, etc. Especially the drop fuse, due to its visible disconnection point and convenient replacement characteristics, is widely used in the medium-voltage distribution network. In recent years, with the development of power system monitoring technology, some improved fuses have begun to integrate simple current detection functions, but most of them are still limited to local display and basic protection functions and cannot achieve remote monitoring and accurate fault location. At the same time, in the field of distribution line fault location, there are already a variety of technical solutions, such as impedance measurement-based positioning methods, fault indicator positioning methods, and traveling wave positioning technologies developed in recent years. However, most of these technologies require special ranging equipment and are independent of the fuse protection function.

[0003] However, traditional fuses and existing fault location technologies have obvious deficiencies. First of all, traditional fuses cannot provide electrical parameter information before and after a fault occurs, nor do they have the ability to judge the fault type and locate the position, resulting in the need for manual line patrol to find faults after the fault occurs, which is time-consuming and inefficient. Secondly, although the existing traveling wave ranging technology has high accuracy, it usually requires expensive special equipment to be installed at both ends of the line, and has extremely high requirements for time synchronization and data processing capabilities, making it difficult to be widely applied at the end of the distribution network. In addition, although the traditional drop fuse can cut off the fault current, it cannot implement a differential protection strategy and adopts the same treatment method for all types of faults, which is neither economical nor conducive to improving the reliability of the power grid. The most crucial thing is that existing technical solutions are difficult to solve the monitoring problem of the "last mile" of the distribution line. Especially in rural power grids and remote areas, due to equipment investment and communication condition restrictions, the coverage of fault location and intelligent monitoring is extremely limited. Summary of the Invention

[0004] The present application provides a traveling wave positioning method, system, device and medium based on an intelligent fuse, which is an intelligent device that can accurately locate the position of the fault point and automatically execute differential protection actions according to the severity of the fault, thereby greatly shortening the fault finding and handling time and improving the operation reliability and maintenance efficiency of the distribution network.

[0005] In a first aspect, the present application provides a traveling wave positioning method based on an intelligent fuse. The traveling wave positioning method based on an intelligent fuse includes: installing intelligent fuses on three phases of a distribution line and performing high-frequency sampling on line signals to obtain digital data containing power frequency signals and high-frequency signals; transmitting the digital data to a master fuse, and using the master fuse to perform zero-sequence quantity synthesis calculation to obtain zero-sequence voltage, zero-sequence current, and a fault judgment result; applying adaptive threshold wavelet transform processing to the collected line signals according to the zero-sequence voltage and the zero-sequence current, extracting multi-layer wavelet coefficients and constructing a feature vector matrix to obtain anti-interference enhanced traveling wave feature information; combining the zero-sequence current and the traveling wave feature information, and using an improved double-ended positioning algorithm to perform ranging calculation on the fault point position to obtain a fault section positioning result; based on the fault judgment result and the fault section positioning result, implementing a hierarchical protection control strategy to obtain predictive maintenance information.

[0006] In a second aspect, the present application provides a traveling wave positioning system based on an intelligent fuse. The traveling wave positioning system based on an intelligent fuse includes: a sampling module, configured to install intelligent fuses on three phases of a distribution line and perform high-frequency sampling on line signals to obtain digital data containing power frequency signals and high-frequency signals; a calculation module, configured to transmit the digital data to a master fuse, and use the master fuse to perform zero-sequence quantity synthesis calculation to obtain zero-sequence voltage, zero-sequence current, and a fault judgment result; an extraction module, configured to apply adaptive threshold wavelet transform processing to the collected line signals according to the zero-sequence voltage and the zero-sequence current, extract multi-layer wavelet coefficients and construct a feature vector matrix to obtain anti-interference enhanced traveling wave feature information; a ranging module, configured to combine the zero-sequence current and the traveling wave feature information, and use an improved double-ended positioning algorithm to perform ranging calculation on the fault point position to obtain a fault section positioning result; a hierarchical module, configured to implement a hierarchical protection control strategy based on the fault judgment result and the fault section positioning result to obtain predictive maintenance information.

[0007] In a third aspect, there is provided a traveling wave positioning device based on an intelligent fuse, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the traveling wave positioning device based on an intelligent fuse executes the above-mentioned traveling wave positioning method based on an intelligent fuse.

[0008] In a fourth aspect, a computer-readable storage medium is provided, in which instructions are stored. When the instructions run on a computer, the computer is enabled to execute the above-mentioned traveling wave positioning method based on an intelligent fuse.

[0009] In the technical solution provided by this application, by installing intelligent fuses on three phases of a distribution line and performing high-frequency sampling on line signals, digital data containing power frequency signals and high-frequency signals is obtained, solving the problem that traditional fuses cannot monitor the line status in real time; the digital data is transmitted to the master fuse and the zero-sequence quantity synthesis calculation is performed to obtain the zero-sequence voltage, zero-sequence current, and fault judgment result, realizing the accurate identification of fault types and improving the accuracy of fault judgment; the adaptive threshold wavelet transform is applied to the collected line signals according to the zero-sequence voltage and zero-sequence current, multi-layer wavelet coefficients are extracted and a feature vector matrix is constructed to obtain anti-interference enhanced traveling wave feature information, greatly improving the anti-interference ability and feature extraction efficiency of traveling wave signal processing; combining the zero-sequence current and traveling wave feature information, an improved double-ended positioning algorithm is used to calculate the distance to the fault point to obtain an accurate fault section positioning result, improving the fault positioning accuracy from hundreds of meters in traditional methods to dozens of meters; based on the fault judgment result and the fault section positioning result, a hierarchical protection control strategy is executed to obtain intelligent fault isolation and predictive maintenance information, realizing the differential control of protection actions and avoiding unnecessary power outages; the present invention makes full use of the key contributions of the artificial intelligence algorithm features to the solution. In particular, through the adaptive threshold wavelet transform processing and the construction of the feature vector matrix, the automatic extraction of effective features in complex power signals is realized. The improved double-ended positioning algorithm greatly improves the accuracy of traveling wave positioning through intelligent parameter adjustment and line impedance correction. The hierarchical protection control strategy uses an intelligent decision-making algorithm to perform differential processing according to the nature of the fault. These algorithm features together constitute the core technical advantages of the present invention, solving the monitoring problem of the "last mile" in the distribution network; in addition, this method also realizes the accurate assessment of the fault influence range and predictive maintenance. Through the time series analysis of historical fault data, the equipment degradation trend and frequently faulty sections are identified, fundamentally improving the operation reliability and maintenance efficiency of the distribution network, and transforming the distribution system from passive protection to active prevention. Through practical application verification, this method can shorten the fault handling time from several hours in traditional methods to dozens of minutes, greatly reducing power outage losses and operation and maintenance costs. Description of the Drawings

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0011] Figure 1 It is a schematic diagram of an embodiment of the traveling wave positioning method based on an intelligent fuse in an embodiment of the present application; Figure 2 It is a schematic diagram of an embodiment of the traveling wave positioning system based on an intelligent fuse in an embodiment of the present application; Figure 3 It is a structural schematic block diagram of the traveling wave positioning device based on an intelligent fuse in an embodiment of the present invention. Specific embodiments

[0012] The embodiments of the present application provide a traveling wave positioning method, system, device and medium based on an intelligent fuse. Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims and above-mentioned drawings of the present application are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" or "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0013] For ease of understanding, the specific process of the embodiments of the present application will be described below. Please refer to Figure 1 , an embodiment of the traveling wave positioning method based on an intelligent fuse in an embodiment of the present application includes: Step S101: Install intelligent fuses on three phases of the distribution line and perform high-frequency sampling on the line signals to obtain digitized data including power frequency signals and high-frequency signals; Step S102: Transmit the digitized data to the master fuse, and use the master fuse to perform zero-sequence quantity synthesis calculation to obtain zero-sequence voltage, zero-sequence current and fault judgment result; Step S103: Apply adaptive threshold wavelet transform processing to the collected line signals according to the zero-sequence voltage and zero-sequence current, extract multi-layer wavelet coefficients and construct a feature vector matrix to obtain anti-interference enhanced traveling wave feature information; Step S104: Combine the zero-sequence current and the traveling wave feature information, and use an improved double-end positioning algorithm to calculate the distance to the fault point to obtain the fault section positioning result; Step S105: Based on the fault judgment result and the fault section positioning result, execute a hierarchical protection control strategy to obtain predictive maintenance information.

[0014] It can be understood that the execution entity of this application can be a traveling wave positioning system based on intelligent fuses, or a terminal or a server. Specifically, it is not limited here. In the embodiments of this application, the server is used as the execution entity for illustration.

[0015] Specifically, intelligent fuses are installed on the three phases of the distribution line and high-frequency sampling is performed on the line signals. The intelligent fuses are equipped with high-frequency current transformers and voltage sensors, and the power line signals are sampled at a sampling frequency of 2 MHz, which greatly exceeds the capabilities of traditional sampling devices. High-frequency sampling can capture both power frequency signals (50 Hz) and high-frequency transient signals (in the range of kilohertz to megahertz) simultaneously, forming complete digital data. For example, when a normal power frequency signal of 5 A rated current flows through the line, the intelligent fuse will capture both this power frequency signal and the possible high-frequency components at the same time. These high-frequency components are particularly obvious in the initial stage of the fault and often contain rich fault characteristic information.

[0016] The digital data is transmitted to the master fuse. The B-phase intelligent fuse is set as the master fuse, and the A-phase and C-phase fuses are set as slave fuses. The slave fuses transmit the collected data to the master fuse through 433 MHz radio frequency communication. After the master fuse receives the three-phase data, it performs zero-sequence quantity synthesis calculation, that is, adding the three-phase voltage vectors and dividing by three to obtain the zero-sequence voltage, and adding the three-phase current vectors and dividing by three to obtain the zero-sequence current. The presence of zero-sequence components is a typical feature of single-phase grounding faults. The master fuse judges the fault according to the magnitude and phase relationship of the zero-sequence voltage and zero-sequence current. When the zero-sequence voltage exceeds the preset threshold and the zero-sequence current exceeds another preset threshold, it is determined as a single-phase grounding fault. For interphase short-circuit faults, the judgment is mainly based on the characteristics of sudden change of phase current and reduction of interphase voltage.

[0017] Adaptive threshold wavelet transform processing is applied to the collected line signals according to the zero-sequence voltage and zero-sequence current. The master fuse dynamically adjusts the processing threshold of the wavelet transform according to the fault severity. The higher the fault severity, the lower the threshold is used to ensure capturing more effective information. Wavelet transform is applied to the original high-frequency data. Usually, the db4 wavelet is selected as the mother wavelet function and decomposed into 8 layers to obtain a series of wavelet coefficients in different frequency bands. Special attention is paid to the first two detail coefficients reflecting high-frequency characteristics, and these coefficients correspond to the main frequency components of the fault traveling wave. The system extracts these wavelet coefficients and constructs a feature vector matrix to form an information set containing multi-dimensional features such as traveling wave amplitude, rise time, spectrum distribution, and waveform polarity.

[0018] Combined with zero-sequence current and traveling wave characteristic information, an improved double-ended location algorithm is used to calculate the distance to the fault point. The double-ended location algorithm is calculated based on the time difference of the traveling wave arriving at two measurement points. When the distance of the fault point in the line is L, the distance between the two measurement points (usually one is the traveling wave location switch on the main line and the other is the intelligent fuse on the branch line) is D, the propagation speed of the electromagnetic wave in the line is V, and the times when the two measurement points capture the traveling wave are T 1 and T 2 , then the calculation formula for the distance S from the fault point to the first measurement point is S = [D - (T 2 - T 1 ) × V] / 2. The improved double-ended location algorithm adds a line impedance correction coefficient to compensate for different types and materials of distribution lines and improve the ranging accuracy.

[0019] Based on the fault judgment result and the fault section location result, a hierarchical protection control strategy is executed. The system classifies the severity of the fault, including emergency faults, serious faults, and minor faults. Different protection actions are taken for different levels of faults. For emergency faults, the fuse drop operation is immediately executed to cut off the current; for serious faults, a delayed drop is executed to leave time for the self-healing of temporary faults; for minor faults, only records are made without execution of cutting off. At the same time, predictive maintenance information is generated. Through time series analysis of historical fault data, the degradation trend of equipment and the frequently fault-occurring sections are identified to form a prediction of possible future faults, guiding maintenance personnel to carry out targeted maintenance.

[0020] For example, during the high-frequency sampling process, an abnormal waveform was captured by the three-phase intelligent fuse on a certain distribution line. The master fuse calculated that the zero-sequence voltage was 6.5 kV (exceeding the 5% threshold of the rated phase voltage of 10 kV), and the zero-sequence current was 2.6 A (exceeding the 10% threshold of the rated current of 5 A), and it was determined as a phase C grounding fault. Then the system performed wavelet transform processing on the high-frequency signal, extracted the traveling wave characteristics from the first two layers of wavelet coefficients, and captured that the time when the traveling wave arrived at the traveling wave location switch on the main line was 10:15:32.000256, and the time when it arrived at the intelligent fuse on the branch line was 10:15:32.000842. Given that the line length between the two devices is 3000 meters and the propagation speed of the electromagnetic wave is 2×10 8 m / s, after calculation, the fault point is 1172 meters away from the traveling wave location switch on the main line. The system determined that this was a serious fault based on the fault, executed a delayed drop action, generated a fault impact assessment report, accurately located the fault area and guided the maintenance personnel to quickly handle it, avoiding a large-scale power outage, and at the same time incorporated the fault data into the historical record for future fault prediction analysis.

[0021] In the embodiment of the present application, by installing intelligent fuses on the three phases of the distribution line and performing high-frequency sampling on the line signals, digital data containing power frequency signals and high-frequency signals is obtained, solving the problem that traditional fuses cannot monitor the line status in real time; the digital data is transmitted to the master fuse and the zero-sequence quantity synthesis calculation is performed to obtain the zero-sequence voltage, zero-sequence current and fault judgment result, realizing the accurate identification of the fault type and improving the accuracy of fault judgment; the adaptive threshold wavelet transform is applied to the collected line signals according to the zero-sequence voltage and zero-sequence current, multi-layer wavelet coefficients are extracted and a feature vector matrix is constructed to obtain anti-interference enhanced traveling wave feature information, greatly improving the anti-interference ability and feature extraction efficiency of traveling wave signal processing; combining the zero-sequence current and traveling wave feature information, an improved double-ended positioning algorithm is used to calculate the distance to the fault point to obtain an accurate fault section positioning result, improving the fault location accuracy from hundreds of meters in the traditional method to dozens of meters; based on the fault judgment result and the fault section positioning result, a hierarchical protection control strategy is executed to obtain intelligent fault isolation and predictive maintenance information, realizing the differential control of protection actions and avoiding unnecessary power outages; the present invention makes full use of the key contributions of the artificial intelligence algorithm features to the solution, especially through the adaptive threshold wavelet transform processing and the construction of the feature vector matrix, realizing the automatic extraction of effective features in complex power signals, and the improved double-ended positioning algorithm greatly improves the accuracy of traveling wave positioning through intelligent parameter adjustment and line impedance correction, and the hierarchical protection control strategy uses an intelligent decision-making algorithm to perform differential processing according to the fault nature. These algorithm features together constitute the core technical advantages of the present invention, solving the monitoring problem of the "last mile" of the distribution network; in addition, this method also realizes the accurate assessment of the fault impact range and predictive maintenance. Through the time series analysis of historical fault data, the equipment degradation trend and frequently faulted sections are identified, fundamentally improving the operation reliability and maintenance efficiency of the distribution network, transforming the distribution system from passive protection to active prevention. Through practical application verification, this method can shorten the fault handling time from several hours in the traditional method to dozens of minutes, greatly reducing the power outage loss and operation and maintenance cost.

[0022] In a specific embodiment, the process of executing step S101 may specifically include the following steps: Install a traveling wave positioning switch on the main line, install intelligent fuses with GPS timing modules on the three phases of the branch line respectively, and perform clock synchronization configuration on the intelligent fuses. The synchronization accuracy of the clock synchronization configuration is 200 nanoseconds; Use the high-frequency current transformer and voltage sensor in the intelligent fuse to collect line signals to obtain the original analog electrical signals; Input the original analog electrical signals into the digital conversion module of the control unit for analog-to-digital conversion processing to obtain digital data; Perform noise suppression and frequency band separation processing on digital data to obtain power frequency signals and high-frequency signals; Calculate the amplitude and phase of the power frequency signal to obtain parameters such as phase voltage, phase current, active power, reactive power, and power factor of the line operating state, and add timestamp information to the high-frequency signal and store it in the local memory.

[0023] Specifically, a monitoring network is established on the distribution line. Traveling wave positioning switches are installed on the main line. The switches are devices for capturing high-frequency fault traveling wave signals and are installed at key nodes of the distribution line such as the substation outlet or the main branch point. On the three phases of the branch line, intelligent fuses are respectively installed. Each intelligent fuse is equipped with a GPS timing module. By receiving GPS satellite signals, the standard time can be obtained with an accuracy of up to the nanosecond level. When performing clock synchronization configuration on the intelligent fuses, a dedicated GPS timing algorithm is used to uniformly calibrate the clocks of the intelligent fuses to the same time reference, ensuring that the clock synchronization accuracy reaches 200 nanoseconds. This high-precision clock synchronization is a key prerequisite for traveling wave positioning because the traveling wave propagation speed is extremely fast. If the clocks are not synchronized, serious deviations will occur in the ranging results.

[0024] The high-frequency current transformer and voltage sensor in the intelligent fuse are the core hardware for obtaining line signals. The high-frequency current transformer uses special magnetic core materials and has broadband response characteristics, capable of capturing both power frequency and high-frequency current signals simultaneously. The voltage sensor uses capacitive voltage division technology to safely convert high voltage into low voltage signals. These two sensors continuously collect line signals to generate raw analog electrical signals. The raw analog electrical signals contain electrical characteristics but have not been digitized and cannot be directly used for computer analysis.

[0025] Input the raw analog electrical signal into the digital conversion module of the control unit. The digital conversion module is mainly composed of a high-speed analog-to-digital converter with a sampling frequency as high as 2 MHz, which is much higher than the sampling frequency of traditional power systems. During the analog-to-digital conversion process, the control unit samples the analog signal at a fixed time interval (0.5 microseconds). Each sampling converts the analog quantity into a digital quantity, forming a series of discrete digital values. These digital values are arranged in chronological order to form a digital data stream.

[0026] Noise suppression uses digital filtering techniques, mainly including mean filtering, median filtering, and low-pass filtering, etc. Mean filtering smooths the data by calculating the average value of the data points within a sliding window; median filtering removes spike noise by taking the median value of the data within the window; low-pass filtering retains the low-frequency components and suppresses high-frequency noise. Band separation uses frequency-domain analysis methods to decompose the signal into different frequency bands. The time-domain signal is converted to the frequency domain through Fourier transform, and then a frequency threshold is set to identify the frequency components near 50Hz as power-frequency signals and the frequency components in the range of several kilohertz to megahertz as high-frequency signals. The inverse transform is performed respectively to obtain the separated power-frequency and high-frequency signals. The power-frequency signal mainly reflects the normal operating state of the line, and amplitude and phase calculations are required. The amplitude calculation uses the root-mean-square method to calculate the square root of the sum of the squares of the sampling points within a power-frequency cycle; the phase calculation is obtained through zero-crossing transition detection or the phase component of Fourier transform. Through these calculations, the phase voltage, phase current, active power, reactive power, and power factor parameters of the line operating state are obtained. The active power is calculated by the average value of the product of the instantaneous values of voltage and current within a cycle; the reactive power is obtained through the analysis of the phase difference between voltage and current; the power factor is the ratio of active power to apparent power. These parameters comprehensively reflect the electrical operating state of the line.

[0027] When processing high-frequency signals, the intelligent fuse adds timestamp information to each sampling data. The timestamp comes from the GPS timing module with an accuracy of 200 nanoseconds. The high-frequency signal data with timestamps is stored in the local memory to form the basic data for traveling wave positioning. The local memory uses high-speed flash memory with sufficient capacity to store high-frequency sampling data in a short time.

[0028] In a specific embodiment, the process of executing step S102 may specifically include the following steps: Send the digitized data to the main-phase intelligent fuse through the radio frequency communication module to obtain a three-phase synchronous data set; Perform timestamp verification processing on the three-phase synchronous data set, screen out the data with time synchronization errors exceeding the preset threshold, and obtain the target three-phase data; Input the target three-phase data into the deep learning transformer model for three-phase voltage vector sum operation, and combine with time-series data prediction to obtain the real-time value of zero-sequence voltage; Input the target three-phase data into the zero-sequence current calculation unit for three-phase current vector sum operation to obtain the real-time value of zero-sequence current; Execute the fault type judgment logic according to the real-time value of zero-sequence voltage and the real-time value of zero-sequence current. When the real-time value of zero-sequence voltage exceeds the first preset threshold and the real-time value of zero-sequence current exceeds the second preset threshold, it is determined as a single-phase ground fault. When the current of any two phases increases and the corresponding phase voltage decreases, it is determined as an interphase short-circuit fault to obtain the fault type recognition result; Based on the fault type recognition result, the phase with the sum of the phase voltage and the zero-sequence voltage less than the original phase voltage is determined as the grounded phase, and the fault judgment result is obtained.

[0029] Specifically, the digital data is sent to the main-phase intelligent fuse via the radio frequency communication module. The radio frequency communication module uses the 433 MHz frequency band, which has good transmission distance and penetration ability. The slave fuses (usually phase A and phase C) pack the digital data collected by each of them into data frames, and the data frames contain information such as sampling data, device identification, sampling timestamp, etc. After adding the CRC check code, the data frames are sent to the master fuse (usually phase B) in a point-to-point manner via the radio frequency communication module. After receiving the data from the three phases, the master fuse unpacks the data and sorts it according to the timestamp to form a three-phase synchronous data set. The timestamp verification process checks the timestamp information in the data frame and aligns the three-phase data according to the timestamp. The preset threshold is set to 200 nanoseconds, and the threshold is determined based on the synchronization accuracy of the GPS timing module. The master fuse will check the timestamp difference of each data point. If it is found that the timestamp of a certain data point differs from the timestamps of the other-phase data at the corresponding moment by more than 200 nanoseconds, then that data point is marked as abnormal data. The method of screening out abnormal data is to replace it with the interpolation of adjacent valid data points or directly discard that data point. The data after the timestamp verification process is the target three-phase data.

[0030] The zero-sequence voltage calculation unit uses a recurrent neural network algorithm. This algorithm constructs a deep learning model containing gated recurrent units (GRUs), which can effectively capture the temporal characteristics and inter-phase relationships of the three-phase voltage data. The input layer of the model receives the three-phase voltage time-series data containing the historical window, extracts the temporal domain features through two layers of bidirectional GRU hidden layers, and then maps them to the output layer through the fully connected layer to obtain the real-time value of the zero-sequence voltage. This model is trained with the historical data of large-scale distribution networks and uses the attention mechanism to automatically identify the voltage characteristics at key time points, and has higher noise immunity and adaptability to non-linear distortion compared with the traditional vector sum operation method. The calculated real-time value of the zero-sequence voltage changes with time, reflecting the dynamic change of the zero-sequence voltage of the line. The process of inputting the target three-phase data into the zero-sequence current calculation unit for three-phase current vector sum operation is similar to that of zero-sequence voltage calculation. The zero-sequence current calculation unit adopts the three-phase current vector sum algorithm, performs vector addition operation on the three-phase currents according to their amplitudes and phases, and then divides the result by 3 to obtain the zero-sequence current. The specific calculation process is as follows: First, represent the three-phase currents as amplitudes and phases, then calculate the sum of the real and imaginary parts of the three-phase currents respectively, divide the sum of the real and imaginary parts by 3 respectively, and finally synthesize the amplitude and phase of the zero-sequence current. The calculated real-time value of the zero-sequence current reflects the dynamic change of the zero-sequence current in the line and is an important basis for judging the fault type. The fault type judgment logic is designed based on the fault characteristics of the power system, and different judgment conditions are formulated for different fault types. For single-phase grounding faults, the judgment condition is that the real-time value of the zero-sequence voltage exceeds the first preset threshold and the real-time value of the zero-sequence current exceeds the second preset threshold. The first preset threshold is usually set to 5% of the rated phase voltage, and the second preset threshold is usually set to 10% of the rated current. When the zero-sequence voltage and zero-sequence current both meet the threshold conditions, it is determined as a single-phase grounding fault. For interphase short-circuit faults, the judgment condition is that the currents of any two phases increase significantly and the corresponding interphase voltages decrease significantly. Interphase short-circuits usually show that the short-circuit phase current reaches several times the normal current, while the interphase voltage drops to several tens of percentages of the normal voltage. By executing the fault type judgment logic, the fault type recognition result is obtained to clearly distinguish whether the fault type is single-phase grounding or interphase short-circuit. Based on the fault type recognition result, the specific fault phase is further determined. For single-phase grounding faults, it is necessary to determine which phase is grounded by comparing the relationship between the vector sum of each phase voltage and the zero-sequence voltage and the original phase voltage. The specific judgment logic is: Calculate the vector sum of each phase voltage and the zero-sequence voltage. If the vector sum of a certain phase is less than the original voltage of that phase, then it is determined that this phase is the grounded phase. This judgment is based on the power system fault theory. In single-phase grounding faults, the voltage of the fault phase is opposite in phase to the zero-sequence voltage, resulting in the vector sum being less than the original phase voltage.

[0031] In a specific embodiment, the process of executing step S103 may specifically include the following steps: Based on the amplitudes of the zero-sequence voltage and zero-sequence current, set a dynamic processing threshold for the line signal to obtain an adaptive signal processing threshold; Input the line signal into the wavelet transform processing unit, and use the mother wavelet function to decompose the signal to obtain eight-layer wavelet decomposition coefficients; Apply the adaptive signal processing threshold to the detail coefficients in the eight-layer wavelet decomposition coefficients for coefficient screening to obtain high-frequency feature coefficients; Input the high-frequency feature coefficients into the sliding window analysis unit, set an analysis window with a fixed duration and a window overlap ratio of 50% to obtain continuous time-series feature analysis segments; Extract four types of characteristic parameters, namely traveling wave amplitude, rise time, spectral distribution, and waveform polarity, from the characteristic analysis segment of continuous time series to obtain a multi-dimensional characteristic parameter set; Organize the multi-dimensional characteristic parameter set into a structured characteristic vector matrix and add timestamp information to obtain anti-interference enhanced traveling wave characteristic information.

[0032] Specifically, the dynamic processing threshold is automatically adjusted according to the severity of the fault to ensure that effective traveling wave signals can be captured under different fault conditions. A mapping relationship between the zero-sequence quantity and the threshold is established. When the zero-sequence voltage amplitude exceeds 5% of the rated phase voltage and the zero-sequence current amplitude exceeds 10% of the rated current, it indicates a relatively high fault severity. At this time, a lower processing threshold is set to capture more traveling wave information; when the zero-sequence quantity is small, a higher threshold is set to filter out more interference signals. The threshold calculation adopts a piecewise linear method. First, the upper and lower limits of the threshold are determined, and then linear interpolation is performed between the upper and lower limits according to the actual value of the zero-sequence quantity. For example, when the rated phase voltage is 10 kV and the rated current is 5 A, if the detected zero-sequence voltage is 1 kV (10% of the rated value) and the zero-sequence current is 1 A (20% of the rated value), the processing threshold is set to 80% of the reference threshold to form an adaptive signal processing threshold.

[0033] Wavelet transform can provide both time-domain and frequency-domain information of signals and is suitable for analyzing non-stationary signals such as fault traveling waves. The wavelet transform processing unit uses the discrete wavelet transform algorithm and selects db4 (Daubechies 4th order) as the mother wavelet function. This wavelet function has good capture ability for spike signals. The discrete wavelet transform process includes two parts: high-pass filtering and low-pass filtering. High-pass filtering obtains detail coefficients (high-frequency information), and low-pass filtering obtains approximation coefficients (low-frequency information). Through eight iterations of decomposition, eight-layer wavelet decomposition coefficients are obtained, including an eighth-layer approximation coefficient and eight detail coefficients (d1 to d8) of different scales. The detail coefficients of the lower layers reflect the high-frequency characteristics of the signal, the detail coefficients of the higher layers reflect the intermediate-frequency characteristics of the signal, and the approximation coefficient reflects the low-frequency characteristics of the signal.

[0034] Apply the adaptive signal processing threshold to the detail coefficients in the eight-layer wavelet decomposition coefficients for coefficient screening. The coefficient screening adopts the hard threshold method. Coefficients smaller than the threshold are set to zero, and coefficients larger than the threshold are retained, thereby removing the influence of noise. The adaptive signal processing threshold will be adjusted when applied to the detail coefficients of different layers. Usually, higher thresholds are used for the detail coefficients of the lower layers (corresponding to high-frequency components), and lower thresholds are used for the detail coefficients of the higher layers (corresponding to intermediate-frequency components). This differential processing takes into account the characteristics of noise distribution in different frequency bands. After coefficient screening, the retained coefficients mainly reflect the characteristics of the fault traveling wave and form high-frequency characteristic coefficients. These high-frequency characteristic coefficients are mainly concentrated in the d1 to d4 layers and contain the main information of the fault traveling wave.

[0035] The sliding window analysis segments the signal using a window of fixed duration. The window length is usually set to 2 milliseconds, and the overlap ratio between adjacent windows is set to 50%, that is, the window moves 1 millisecond each time. The sliding window moves step by step on the time axis, analyzes the high-frequency feature coefficients within each window, and forms a series of feature analysis segments. Each feature analysis segment is associated with a specific time point, marking the specific moment when the segment is located.

[0036] The traveling wave amplitude parameter is obtained by calculating the maximum peak value of the signal within the window, reflecting the intensity of the traveling wave; the rise time parameter is obtained by calculating the time required for the signal to rise from 10% of the peak value to 90% of the peak value, reflecting the steepness of the traveling wave; the spectrum distribution parameter is obtained by performing a fast Fourier transform on the signal within the window and calculating the energy ratio of different frequency bands, reflecting the frequency domain characteristics of the traveling wave; the waveform polarity parameter is obtained by judging the positive and negative directions of the initial waveform, reflecting the orientation of the fault point relative to the measurement point. These four types of feature parameters together constitute a multi-dimensional feature parameter set, comprehensively describing the characteristics of the traveling wave signal. Each parameter has its physical meaning: the larger the amplitude, the more serious the fault; the shorter the rise time, the steeper the front of the traveling wave; the more high-frequency components in the spectrum distribution, the sharper the traveling wave signal; and the waveform polarity is directly related to the judgment of the fault direction.

[0037] The feature vector matrix adopts a two-dimensional structure with the row being the time series and the column being different feature parameters. Each matrix element corresponds to the value of a specific feature parameter at a specific moment. The matrix also contains accurate timestamp information, marking the sampling moment corresponding to each row of data. These timestamps are from the GPS timing module with an accuracy of 200 nanoseconds. The structured feature vector matrix is convenient for subsequent algorithm processing, especially for the application of the traveling wave positioning algorithm.

[0038] In a specific embodiment, the process of executing step S104 may specifically include the following steps: Judge the line range where the fault section is located according to the zero-sequence current. When the zero-sequence current exceeds the first preset threshold, select the traveling wave positioning switch of the main line and the intelligent fuse of the branch line as the measurement points. When the zero-sequence current is lower than the first preset threshold but higher than the second preset threshold, select two adjacent intelligent fuses as the measurement points to obtain the traveling wave information of the two ends of the measurement points participating in the positioning calculation; Extract the traveling wave arrival timestamps from the traveling wave information of the two ends of the measurement points, calculate the difference between the traveling wave arrival times at the two ends, and obtain the propagation time difference of the fault traveling wave; Multiply the propagation time difference of the fault traveling wave by the propagation speed value of the electromagnetic wave in the line, and subtract the obtained product result from the line length between the two ends of the measurement points to obtain twice the fault distance value; Divide twice the fault distance value by two to obtain the target distance value from the fault point to the first-end measurement point; Perform line topology matching on the target distance value. When the fault is located at a T-junction, use the traveling wave reflection feature for secondary verification. When the fault is located on a straight-line line, directly confirm the position to obtain the preliminary fault section location result; Compare the preliminary fault section location result with the historical fault database, extract the deviation value between the actual position and the calculated position of the historical similar fault, and adjust the deviation value according to the current line load rate and seasonal factors to obtain the fault section location result.

[0039] Specifically, divide the section by the magnitude of the zero-sequence current. The zero-sequence current judgment adopts a double-threshold mechanism. The first preset threshold is usually set to 15% of the rated current, and the second preset threshold is set to 5% of the rated current. When the master fuse detects that the zero-sequence current exceeds the first preset threshold, it indicates that the fault area is relatively clear and the fault is relatively serious. At this time, select the traveling wave positioning switch on the main line and the intelligent fuse on the branch line as the measurement point combination. This combination is applicable to the situation where the fault occurs in the connection area between the main line and the branch. When the zero-sequence current is lower than the first preset threshold but higher than the second preset threshold, it indicates that the fault is relatively light or the position is relatively far. At this time, select two adjacent intelligent fuses as the measurement point combination. This combination is applicable to the situation where the fault occurs at the end of the branch line. Through this adaptive selection mechanism, the two-end measurement points most suitable for the current fault situation are determined, and complete traveling wave information, including traveling wave waveform, amplitude, phase, and accurate timestamp and other data, is extracted from these two-end measurement points.

[0040] Extract the traveling wave arrival timestamp from the traveling wave information of the two-end measurement points. The traveling wave arrival timestamp refers to the exact moment when the fault traveling wave first arrives at the measurement point, which is determined by the traveling wave characteristic information extracted by wavelet transform. The traveling wave arrival judgment adopts the threshold detection method, and the moment when the amplitude of the characteristic signal first exceeds the preset threshold is recorded as the traveling wave arrival time. The two-end measurement points respectively record their own traveling wave arrival times, and then calculate the difference between these two times to obtain the traveling wave propagation time difference of the fault traveling wave.

[0041] Multiply the traveling wave propagation time difference of the fault traveling wave by the propagation speed value of the electromagnetic wave in the line, and then subtract the obtained product result from the line length between the two-end measurement points. The propagation speed value of the electromagnetic wave in the power line depends on the line type. For overhead lines, it is usually taken as 2 / 3 of the speed of light; for cable lines, it is taken as 1 / 3 to 1 / 2 of the speed of light. The line length between the two-end measurement points refers to the distance measured along the actual path of the line, and the bending and turning of the line need to be considered. When calculating, subtract the product of the time difference and the speed from the line length, and the obtained result is the sum of the distances from the fault point to the two measurement points, that is, twice the fault distance value.

[0042] Divide the double fault distance value by two to obtain the target distance value from the fault point to the head-end measurement point. This step converts the calculation result into the form required for actual engineering applications. The head-end measurement point usually refers to the measurement point where the traveling wave arrives first, or a pre-specified reference measurement point. The target distance value directly represents the distance from the fault point along the actual path of the line to the head-end measurement point. Perform line topology matching on the target distance value. Line topology matching needs to combine the actual structure of the distribution line to map the calculated distance value onto the actual line. When the fault is located at a T-junction, due to the complex reflection and transmission phenomena of the traveling wave at the branch point, relying solely on distance calculation may lead to misjudgment. At this time, the traveling wave reflection characteristics need to be used for secondary verification. The traveling wave reflection characteristics refer to the waveform characteristics of the traveling wave reflected back to the measurement point at the branch point or the fault point, including the polarity, amplitude, and arrival time of the reflected wave, etc. By analyzing these characteristics, it can be determined whether the fault occurs at the branch point or on the straight section. When the fault is located on a straight-line type line, the traveling wave propagation path is clear, and the position can be directly confirmed without additional verification. The result of line topology matching forms the preliminary fault section location result, clearly indicating the specific line segment where the fault is located.

[0043] Compare the preliminary fault section location result with the historical fault database. The historical fault database records the past fault information, including fault type, fault location, calculated location, actual location, environmental conditions, etc. Through database retrieval, find historical fault cases similar to the current fault type and location, and extract the deviation values between the calculated location and the actual location in these cases. The deviation value reflects the systematic error of the location algorithm under specific conditions, which is caused by various factors such as line parameter changes and measurement errors. The main control fuse will dynamically adjust the deviation value according to the current line load rate and seasonal factors. The line load rate will affect the line temperature, which in turn affects the electromagnetic wave propagation speed; seasonal factors such as temperature and humidity will also affect the line characteristics. Apply the adjusted deviation value to the preliminary location result to obtain a more accurate fault section location result.

[0044] In a specific embodiment, the process of executing step S105 may specifically include the following steps: Evaluate the severity of the fault judgment result. Divide the fault into three levels: emergency fault, serious fault, and minor fault according to the amplitude and duration of the zero-sequence voltage and zero-sequence current to obtain the fault severity classification result; Formulate a differential protection action plan according to the fault severity classification result and the fault section location result. Immediately execute the dropout action for the emergency fault, execute the delayed dropout action for the serious fault, and only record without executing the dropout for the minor fault to obtain the intelligent fuse dropout control instruction; Transmit the intelligent fuse dropout control instruction to the corresponding intelligent fuse control unit to trigger the dropout mechanism of the intelligent fuse to perform the opening operation, and obtain the physical isolation effect of the fault section; Conduct a correlation analysis on the fault section location result and the distribution system topology structure to delimit the affected user scope and key load conditions, and obtain the fault impact assessment report; Upload the fault judgment result, fault section location result and fault impact assessment report to the main station of the distribution automation system through the 4G communication module, generate a maintenance work order including the longitude and latitude coordinates of the fault point, fault type, fault time and recommended treatment plan, and obtain the dispatching information for the operation and maintenance personnel; Conduct time series analysis on the historical fault data to identify the equipment degradation trend and frequently faulted sections, and obtain the predictive maintenance information.

[0045] Specifically, set the basic score according to the zero-sequence voltage amplitude. When the zero-sequence voltage exceeds 20% of the rated phase voltage, the basic score is 10 points; when it exceeds 10% but is lower than 20%, the basic score is 5 points; when it is lower than 10% but higher than 5%, the basic score is 2 points. Then, correct the basic score according to the zero-sequence current amplitude. When the zero-sequence current exceeds 30% of the rated current, the basic score is increased by 5 points; when it exceeds 15% but is lower than 30%, it is increased by 3 points; when it is lower than 15% but higher than 5%, it is increased by 1 point. Conduct a secondary correction according to the fault duration. If the duration exceeds 5 seconds, the basic score is increased by 3 points; if it exceeds 1 second but is lower than 5 seconds, it is increased by 1 point; if it is lower than 1 second, no score is added. Divide the faults into three levels according to the final score: a score greater than or equal to 12 is an emergency fault, a score between 6 and 11 is a serious fault, and a score of 5 or less is a minor fault.

[0046] Differentiated protection action plans are formulated based on the fault severity classification results and fault section location results, and intelligent decision-making logic is adopted. For emergency faults, the dropout action is immediately executed without waiting or secondary confirmation, directly triggering the fuse dropout mechanism. Emergency faults are usually faults that pose a great threat to equipment, such as severe short circuits or three-phase unbalances. For severe faults, a delayed dropout action is executed, and the set delay time is usually 0.5 to 2 seconds. During this period, the fault status is continuously monitored. If the fault is eliminated by itself, the dropout command is cancelled. If the fault persists, the dropout is executed after the delay ends. For minor faults, only records are made without executing the dropout. Only the fault information is stored and reported, and no power-off operation is performed. The differentiated protection action plan is automatically generated by the master fuse according to the preset logic, forming an accurate intelligent fuse dropout control command. The command includes the action type (immediate dropout, delayed dropout or no dropout), the target fuse identifier and the execution time. For emergency and severe faults, the master fuse sends the dropout control command to the target fuse through 433MHz radio frequency communication. After receiving the command, the control unit of the target fuse first verifies the command to confirm the integrity and validity of the command, and then activates the fuse dropout mechanism. The dropout mechanism is an electromagnetic release device that releases the mechanical lock through the magnetic force generated by the energization of the electromagnetic coil, causing the fuse tube to drop under the action of gravity, realizing the physical disconnection of the circuit. During the dropout process, the position sensor continuously monitors the position status of the fuse tube. When the full dropout signal is detected, the control unit generates a disconnection confirmation message and feeds it back to the master fuse through radio frequency communication, forming a closed-loop control to ensure that the fault section is physically isolated.

[0047] Correlation analysis is carried out on the fault section location results and the distribution system topology structure. The correlation analysis first extracts the network structure information related to the fault section from the distribution network topology database to determine the electrical connection relationship upstream and downstream of the fault point. Then, the affected distribution transformer substations are identified, and the specific scope affected by the power outage is determined by tracing all the distribution transformers downstream of the fault point. For each affected distribution transformer substation, the user information within the substation is further extracted, including the user type (residential, industrial and commercial, agricultural or special users) and the user importance level. In particular, important loads such as hospitals, schools, and government agencies are marked to determine whether there is an emergency power supply demand. These information are integrated to generate a fault impact assessment report, which details the number of affected substations, the number of users, the critical load situation, and the estimated economic losses, etc.

[0048] Uploading the fault judgment results, fault section location results and fault impact assessment reports to the distribution automation system master station through the 4G communication module is a key step in achieving information sharing. The 4G communication module built into the master control fuse uses the mobile communication network to package the information into a data frame in a standard format and send it to the distribution automation system master station through an encrypted channel. After receiving the information, the master station stores it in the fault management database and marks the latitude and longitude coordinates of the fault point on the electronic map through GIS (Geographic Information System). The system automatically generates a maintenance work order based on the fault type, fault location and impact range. The work order contains the precise location description of the fault point, fault type, fault occurrence time, fault severity, impact range and recommended processing solutions. After the work order is generated, it is pushed to the nearest operation and maintenance personnel through SMS, mobile applications, etc., so that the operation and maintenance personnel can distribute information and guide on-site emergency repair work.

[0049] Time series analysis of all previous fault data is the core of predictive maintenance. Time series analysis uses sliding window technology to perform statistical and trend analysis on fault data within a certain period of time in the past (such as 6 months or 1 year). The analysis process classifies fault data according to line segment, fault type, environmental conditions and other dimensions, and then calculates the frequency, interval time and severity change trend of various faults. The failure probability of each line segment in the future period is predicted through algorithms such as linear regression or exponential smoothing. For line segments with a significant increase in fault frequency, further analyze the cause of the fault and identify possible signs of equipment degradation, such as an increase in the number of fuses and aggravated zero-sequence current fluctuations. Combined with information such as equipment installation time, operating environment, and load conditions, an equipment degradation assessment model is formed to predict the remaining life of the equipment. Finally, predictive maintenance information is generated, including a list of high-risk line segments, recommended maintenance time, and key inspection items, providing data support for planned maintenance.

[0050] For example: Under the conditions of winter rain and snow weather on a certain 10kV distribution line, the B-phase main control intelligent fuse captures a zero-sequence voltage of 2.5kV (25% of the rated phase voltage of 10kV), a zero-sequence current of 2.3A (46% of the rated current of 5A), and the fault duration is 3 seconds. According to the scoring mechanism calculation: the basic score of the zero-sequence voltage is 10 points, the zero-sequence current correction increases by 5 points, and the duration correction increases by 1 point, with a total score of 16 points, which is determined to be an emergency fault. At the same time, the traveling wave positioning algorithm calculates that the fault point is located on a section of mountainous line 2.3 kilometers away from the substation. The main control fuse immediately generates a drop control instruction and sends it to the intelligent fuse corresponding to the fault phase through 433MHz radio frequency communication, triggering the drop mechanism to act and completing the power-off operation within 0.1 second after the fault occurs. The position sensor confirms the completion of the drop and feedbacks the status. Subsequently, the system conducts a correlation analysis on the fault section, determines that 3 distribution transformer substations are affected, a total of 120 users, including 1 township health center (belonging to important loads). All information is uploaded to the distribution automation master station through the 4G module, and the system automatically generates a maintenance work order, pushing the precise location of the fault point (longitude 119.xx, latitude 30.xx), fault type (single-phase grounding), occurrence time (2023-12-15 08:23:45), and treatment suggestions to the regional operation and maintenance personnel. At the same time, time series analysis finds that 4 similar faults have occurred in this section in the past 6 months, with a frequency significantly higher than other sections. The prediction analysis shows that the insulators of this line may have aging problems, and it is recommended to conduct a comprehensive inspection and necessary replacement of the insulators of the entire line while dealing with this fault. According to the precise positioning and treatment suggestions, the operation and maintenance personnel directly go to the fault point and find that an insulator is damaged due to excessive rain and snow load, resulting in single-phase grounding caused by the contact between the phase wire and the tower. The entire fault handling process takes only 1.5 hours from occurrence to repair completion, saving more than 3 hours compared with the traditional search method.

[0051] In a specific embodiment, the process of performing the step of correlating the fault section positioning result with the distribution system topology structure may specifically include the following steps: Perform a spatial mapping of the fault section positioning result and the distribution network topology data to determine the line section where the fault point is located and the electrical connection relationship, and obtain the position of the network topology breakpoint; Identify all distribution nodes that cannot obtain power supply according to the network topology breakpoint position to obtain the boundary of the precise power outage range; Perform a clustering analysis on the distribution transformer substation data within the boundary of the precise power outage range, and stratify according to the power supply capacity and load characteristics to obtain hierarchical load impact data; Extract the information of first-level guaranteed users from the hierarchical load impact data, and combine it with the standby power supply configuration status to obtain the list of emergency power supply requirements; Calculate the load loss quantification index based on the precise power outage scope boundary and hierarchical load impact data, apply the time-weighted factor to establish an economic loss model, and obtain the severity curve of the impact in different time periods; Integrate the network topology breakpoint location, precise power outage scope boundary, hierarchical load impact data, emergency power supply demand list, and severity curve of the impact in different time periods into a hierarchical structure report to obtain a fault impact assessment report.

[0052] Specifically, spatially map the fault section location result and the distribution network topology data. In the spatial mapping process, use the line geometric information and electrical connection relationship pre-stored in the distribution network topology database to convert the fault point distance value calculated by traveling wave location into specific physical position coordinates. When specifically implemented, extract the section information of the fault line from the distribution network topology database, including the starting coordinates, ending coordinates, length, and connection relationship of each section of the line. Then locate the distance value from the fault point to the reference point on the line to determine the specific line segment where the fault point is located and its precise position on that line segment. This process takes into account the actual direction and curvature of the line, and by accumulating the lengths of each line segment until the accumulated value exceeds or is equal to the fault distance value, the fault point location is determined. At the same time, extract the electrical connection information of the line section where the fault point is located, including the positions and operating states of upstream and downstream switches, fuses, tapping boxes, etc., draw a local electrical connection diagram, and finally determine the network topology breakpoint location, that is, the power grid disconnection point caused by the fault. Identify all distribution nodes that cannot obtain power supply according to the network topology breakpoint location to determine the power outage scope. Use the electrical connectivity analysis algorithm to trace all nodes that lose power supply downstream from the network topology breakpoint location. The algorithm starts from the breakpoint and traverses all downstream branches of the network diagram, marking all nodes that lose connection with the power supply. When specifically implemented, first establish a directed graph model of the distribution network, where nodes represent devices such as substations, distribution transformer stations, and tapping boxes, and edges represent line connection relationships. Then start from the power supply node and perform a breadth-first search, marking all reachable nodes as energized. Remove the network topology breakpoint from the graph, and perform a breadth-first search again from the power supply node. All unreachable nodes are power-off nodes. By comparing the differences in the results of the two searches, accurately determine the set of nodes powered off due to this fault to form a precise power outage scope boundary. This boundary is stored in two forms: a node list and a geographical coordinate set.

[0053] Cluster analysis first extracts the characteristic data of all distribution transformer substations within the power outage range, including information such as transformer capacity, number of users, user type distribution, peak-valley load difference, etc. Then, the K-means clustering algorithm or hierarchical clustering algorithm is used to classify the substations into different levels according to the power supply capacity and load characteristics. The power supply capacity classification is usually based on the rated capacity of the transformer. For example, less than 100 kVA is small capacity, 100 - 315 kVA is medium capacity, and more than 315 kVA is large capacity. The load characteristic classification considers factors such as the peak-valley difference of the load curve, power factor, and seasonal changes, and classifies the load into categories such as stable type, fluctuating type, and peak type. In the clustering process, the Euclidean distance or cosine similarity between the characteristic vectors of each substation is calculated, and the substations with similar characteristics are grouped into the same category. The finally formed classified load impact data includes information such as the number of substations, total capacity, number of users, and typical load curves for each level, providing a data basis for the fault impact assessment. First-level guaranteed users refer to special users with extremely high requirements for power supply reliability, such as hospitals, schools, government agencies, communication base stations, etc. In the extraction process, all user information within the power outage range is first queried from the user database, and according to the preset user importance classification standard, the users belonging to the first-level guaranteed category are screened out. Then, these users are further analyzed to extract key information such as their electricity consumption capacity, key load description, and maximum allowable power outage time. At the same time, the backup power supply configuration status of these users is queried, including the capacity and sustainable power supply time of self-provided generators, UPS systems, or dual-power switching devices. By comparing the key load requirements of users with the backup power supply capacity, the actual impact degree of the power outage on these users is evaluated to determine whether additional emergency power supply measures are needed. The evaluation results form an emergency power supply demand list, including the list of users who need to have their power supply restored first or need mobile power generation vehicle support, as well as the specific requirements and contact information of each user.

[0054] Based on the accurate power outage range boundary and classified load impact data, the load loss quantification index is calculated. The load loss quantification index includes three dimensions: power outage capacity, power outage electricity, and economic loss. The power outage capacity is directly obtained by accumulating from the classified load impact data, representing the total transformer capacity that has its power supply interrupted due to the fault. The power outage electricity needs to be calculated in combination with the typical load curve. The method is to multiply the typical load factor of each type of substation at different times by the transformer capacity and then by the power outage duration to obtain the estimated power outage electricity. For the economic loss calculation, a time-weighted factor is introduced. The economic losses caused by power outages at different times vary greatly. For example, a power outage during the day on a weekday has a much greater impact on industrial and commercial users than a power outage at night. The time-weighted factor is obtained through historical data statistics and is usually divided into categories such as daytime on weekdays, nighttime on weekdays, weekends, and holidays. Different unit electricity economic loss coefficients are set for each type of time period. Multiply the power outage electricity by the economic loss coefficient of the corresponding time period to obtain the estimated economic loss value for each time period, forming a severity curve of impact by time period, visually showing the change trend of the fault impact over time.

[0055] Integrating the network topology breakpoint location, precise power outage scope boundary, hierarchical load impact data, emergency power supply demand list, and time-segmented impact severity curve into a hierarchical structure report is a key step in forming the final fault impact assessment report. The integration process adopts a hierarchical data organization method, arranging various types of information according to spatial and logical relationships. First is the fault overview section, including the basic information of the fault occurrence time, type, and network topology breakpoint location; second is the power outage scope section, including the text description and map display of the precise power outage scope boundary; then is the user impact section, containing the statistical tables and charts of the hierarchical load impact data; next is the key concern section, listing the key user information in the emergency power supply demand list; finally is the economic assessment section, showing the time-segmented impact severity curve and the overall economic loss estimate. Standardized data formats and unified information coding are adopted during integration to ensure the consistency and readability of the report. At the same time, a unique identification code is generated for the report to facilitate subsequent query and statistical analysis. The finally formed fault impact assessment report contains both detailed technical data and clear management decision-making reference value.

[0056] The above describes the traveling wave positioning method based on intelligent fuses in the embodiments of the present application. Next, the traveling wave positioning system based on intelligent fuses in the embodiments of the present application will be described. Please refer to Figure 2 , an embodiment of the traveling wave positioning system based on intelligent fuses in the embodiments of the present application includes: A sampling module 201, configured to install intelligent fuses on three phases of a distribution line and perform high-frequency sampling on line signals to obtain digitized data containing power frequency signals and high-frequency signals; A calculation module 202, configured to transmit the digitized data to a master fuse, and use the master fuse to perform zero-sequence quantity synthesis calculation to obtain zero-sequence voltage, zero-sequence current, and a fault judgment result; An extraction module 203, configured to apply adaptive threshold wavelet transform processing to the collected line signals according to the zero-sequence voltage and the zero-sequence current, extract multi-layer wavelet coefficients, and construct a feature vector matrix to obtain anti-interference enhanced traveling wave feature information; A ranging module 204, configured to combine the zero-sequence current and the traveling wave feature information, and use an improved double-end positioning algorithm to perform ranging calculation on the fault point location to obtain a fault section positioning result; A grading module 205, configured to execute a grading protection control strategy based on the fault judgment result and the fault section positioning result to obtain predictive maintenance information.

[0057] Through the collaborative cooperation of the above-mentioned various components, by installing intelligent fuses on the three phases of the distribution line and performing high-frequency sampling on the line signals, digital data containing power frequency signals and high-frequency signals is obtained, solving the problem that traditional fuses cannot monitor the line status in real time; the digital data is transmitted to the master fuse and the zero-sequence quantity synthesis calculation is performed to obtain the zero-sequence voltage, zero-sequence current and fault judgment result, realizing the accurate identification of the fault type and improving the accuracy of fault judgment; the adaptive threshold wavelet transform is applied to the collected line signals according to the zero-sequence voltage and zero-sequence current, multi-layer wavelet coefficients are extracted and a feature vector matrix is constructed to obtain anti-interference enhanced traveling wave feature information, greatly improving the anti-interference ability and feature extraction efficiency of traveling wave signal processing; combining the zero-sequence current and traveling wave feature information, an improved double-ended location algorithm is used to calculate the distance to the fault point to obtain an accurate fault section location result, improving the fault location accuracy from hundreds of meters of the traditional method to dozens of meters; based on the fault judgment result and the fault section location result, a hierarchical protection control strategy is executed to obtain intelligent fault isolation and predictive maintenance information, realizing the differential control of protection actions and avoiding unnecessary power outages; the present invention makes full use of the key contributions of the artificial intelligence algorithm features to the solution, especially through the adaptive threshold wavelet transform processing and the construction of the feature vector matrix, realizing the automatic extraction of effective features in complex power signals. The improved double-ended location algorithm greatly improves the accuracy of traveling wave location through intelligent parameter adjustment and line impedance correction. The hierarchical protection control strategy uses an intelligent decision-making algorithm to perform differential processing according to the nature of the fault. These algorithm features together constitute the core technical advantages of the present invention, solving the monitoring problem of the "last mile" of the distribution network; in addition, this method also realizes the accurate assessment of the fault influence range and predictive maintenance. By performing time series analysis on historical fault data, the equipment degradation trend and frequently fault-occurring sections are identified, fundamentally improving the operation reliability and maintenance efficiency of the distribution network, and transforming the distribution system from passive protection to active prevention. Through actual application verification, this method can shorten the fault handling time from several hours of the traditional method to dozens of minutes, greatly reducing the power outage loss and operation and maintenance cost.

[0058] Above Figure 2 From the perspective of modular functional entities, the traveling wave location system based on intelligent fuses in the embodiments of the present invention is described in detail. Next, the traveling wave location device based on intelligent fuses in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0059] Figure 3FIG. 0 is a schematic structural diagram of a traveling wave positioning device based on an intelligent fuse provided by an embodiment of the present invention. The traveling wave positioning device 300 based on the intelligent fuse may vary greatly due to different configurations or performances, and may include one or more processors (central processing units, CPUs) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 for storing application programs 333 or data 332 (for example, one or more mass storage device terminals). Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the traveling wave positioning device 300 based on the intelligent fuse. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of instruction operations in the storage media 330 on the traveling wave positioning device 300 to implement the steps of the above-mentioned traveling wave positioning method based on the intelligent fuse.

[0060] The traveling wave positioning device 300 based on the intelligent fuse may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that Figure 3 The shown structural diagram of the traveling wave positioning device based on the intelligent fuse does not limit the traveling wave positioning device based on the intelligent fuse provided by the present invention, and may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0061] The present invention also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium, or may also be a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer is caused to execute the steps of the traveling wave positioning method based on the intelligent fuse.

[0062] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, systems, and units may refer to the corresponding processes in the foregoing method embodiments and will not be described herein again.

[0063] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a traveling wave positioning device based on an intelligent fuse (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0064] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of various embodiments of the present invention.

Claims

1. A traveling wave positioning method based on intelligent fuse, characterized in that: The method comprises: Install intelligent fuses on the three phases of the distribution line and perform high-frequency sampling on the line signal to obtain digital data containing power frequency signals and high-frequency signals; The digital data is transmitted to a main control fuse, and the main control fuse is used to perform zero-sequence quantity synthesis calculation to obtain zero-sequence voltage, zero-sequence current and fault judgment results; Applying adaptive threshold wavelet transform processing to the collected line signal according to the zero-sequence voltage and the zero-sequence current, extracting multi-layer wavelet coefficients and constructing a feature vector matrix to obtain anti-interference enhanced traveling wave feature information; Combining the zero-sequence current and the traveling wave characteristic information, an improved double-terminal positioning algorithm is used to perform distance calculation on the fault point position to obtain a fault section positioning result; Based on the fault judgment result and the fault section location result, a hierarchical protection control strategy is executed to obtain predictive maintenance information.

2. The traveling wave positioning method based on the intelligent fuse according to claim 1 is characterized in that: The method of installing intelligent fuses on the three phases of the power distribution line and performing high frequency sampling on the line signal to obtain digital data including power frequency signals and high frequency signals includes: Install a traveling wave positioning switch on the trunk line, install smart fuses with GPS timing modules on the three phases of the branch line, and perform clock synchronization configuration on the smart fuses. The synchronization accuracy of the clock synchronization configuration is 200 nanoseconds. The high-frequency current transformer and voltage sensor in the intelligent fuse are used to collect data on the line signal to obtain the original analog electrical signal; Inputting the original analog electrical signal into the digital conversion module of the control unit for analog-to-digital conversion processing to obtain digital data; Performing noise suppression and frequency band separation processing on the digitized data to obtain an industrial frequency signal and a high frequency signal; The amplitude and phase of the power frequency signal are calculated to obtain the phase voltage, phase current, active power, reactive power and power factor parameters of the line operation status, and timestamp information is added to the high frequency signal and stored in the local memory.

3. The traveling wave positioning method based on the intelligent fuse according to claim 1 is characterized in that: The method of transmitting the digital data to a main control fuse and using the main control fuse to perform zero-sequence quantity synthesis calculation to obtain zero-sequence voltage, zero-sequence current and fault judgment results includes: The digital data is sent to the main phase intelligent fuse through the radio frequency communication module to obtain a three-phase synchronous data set; Performing timestamp verification processing on the three-phase synchronous data set, filtering out data whose time synchronization error exceeds a preset threshold, and obtaining target three-phase data; The target three-phase data is input into the deep learning transformer model to perform three-phase voltage vector sum operation, and combined with the time series data prediction to obtain the real-time value of the zero-sequence voltage; Inputting the target three-phase data into a zero-sequence current calculation unit to perform a three-phase current vector sum operation to obtain a real-time value of the zero-sequence current; Execute the fault type judgment logic according to the real-time value of the zero-sequence voltage and the real-time value of the zero-sequence current, determine it as a single-phase grounding fault when the real-time value of the zero-sequence voltage exceeds a first preset threshold and the real-time value of the zero-sequence current exceeds a second preset threshold, and determine it as a phase-to-phase short circuit fault when the currents of any two phases increase and the corresponding phase-to-phase voltage decreases, and obtain a fault type identification result; Based on the fault type identification result, the phase whose phase voltage and zero-sequence voltage vector sum is less than the original phase voltage is determined to be a grounded phase to obtain a fault judgment result.

4. The traveling wave positioning method based on the intelligent fuse according to claim 1 is characterized in that: The method of applying adaptive threshold wavelet transform processing to the collected line signal according to the zero-sequence voltage and the zero-sequence current, extracting multi-layer wavelet coefficients and constructing a feature vector matrix to obtain anti-interference enhanced traveling wave feature information includes: Setting a dynamic processing threshold for the line signal based on the amplitude of the zero-sequence voltage and the zero-sequence current to obtain an adaptive signal processing threshold; The line signal is input into the wavelet transform processing unit, and the mother wavelet function is used to decompose the signal to obtain eight layers of wavelet decomposition coefficients; Applying the adaptive signal processing threshold to the detail coefficients in the eight-layer wavelet decomposition coefficients to perform coefficient screening to obtain high-frequency characteristic coefficients; Inputting the high-frequency characteristic coefficient into a sliding window analysis unit, setting an analysis window of fixed length and a window overlap ratio of 50%, and obtaining a characteristic analysis segment of a continuous time series; Extracting four types of characteristic parameters, namely, traveling wave amplitude, rise time, spectrum distribution and waveform polarity, from the characteristic analysis segment of the continuous time series to obtain a multi-dimensional characteristic parameter set; The multi-dimensional characteristic parameter set is organized into a structured characteristic vector matrix, and timestamp information is added to obtain anti-interference enhanced traveling wave characteristic information.

5. The traveling wave positioning method based on the intelligent fuse according to claim 1 is characterized in that: The method combines the zero-sequence current and the traveling wave characteristic information, adopts an improved double-terminal positioning algorithm to perform distance calculation on the fault point position, and obtains the fault section positioning result, including: The line range where the fault section is located is determined according to the zero-sequence current. When the zero-sequence current exceeds a first preset threshold, a main line traveling wave positioning switch and a branch line intelligent fuse are selected as measurement points. When the zero-sequence current is lower than the first preset threshold but higher than a second preset threshold, two adjacent intelligent fuses are selected as measurement points to obtain traveling wave information of the two end measurement points involved in the positioning calculation. Extracting the traveling wave arrival timestamp from the traveling wave information of the two end measurement points, calculating the difference of the traveling wave arrival times at the two ends, and obtaining the fault traveling wave propagation time difference; Multiply the fault traveling wave propagation time difference by the propagation speed of the electromagnetic wave in the line, and subtract the obtained product from the line length between the measuring points at both ends to obtain twice the fault distance value; Divide the twice fault distance value by two to obtain a target distance value from the fault point to the head-end measurement point; The target distance value is matched with the line topology. When the fault is located at the T-type connection, the traveling wave reflection feature is used for secondary verification. When the fault is located at the straight line, the position is directly confirmed to obtain the preliminary fault section location result. The preliminary fault section location result is compared with the historical fault database, the deviation value between the actual location and the calculated location of similar historical faults is extracted, and the deviation value is adjusted according to the current line load rate and seasonal factors to obtain the fault section location result.

6. The traveling wave positioning method based on the intelligent fuse according to claim 1 is characterized in that: The step of executing a hierarchical protection control strategy based on the fault judgment result and the fault section location result to obtain predictive maintenance information includes: The fault severity is evaluated on the fault judgment result, and the fault is divided into three levels: emergency fault, serious fault and minor fault according to the amplitude and duration of zero-sequence voltage and zero-sequence current, so as to obtain a fault severity classification result; A differentiated protection action plan is formulated according to the fault severity classification result and the fault section location result, and a drop action is immediately executed for an emergency fault, a delayed drop action is executed for a serious fault, and only a record is made for a minor fault without executing the drop, so as to obtain an intelligent fuse drop control instruction; The intelligent fuse drop control instruction is transmitted to the corresponding intelligent fuse control unit, triggering the drop mechanism of the intelligent fuse to perform a breaking operation, thereby obtaining a physical isolation effect of the fault section; Conduct correlation analysis on the fault section location result and the topological structure of the power distribution system, define the affected user range and key load conditions, and obtain a fault impact assessment report; The fault judgment result, the fault section location result and the fault impact assessment report are uploaded to the main station of the distribution automation system through the 4G communication module, and a maintenance work order including the longitude and latitude coordinates of the fault point, the fault type, the fault time and the recommended treatment plan is generated, and the dispatch information of the operation and maintenance personnel is obtained; Conduct time series analysis on all previous failure data to identify equipment degradation trends and failure-frequently occurring sections, and obtain predictive maintenance information.

7. The traveling wave positioning method based on the intelligent fuse according to claim 6 is characterized in that: The fault section location result is correlated with the distribution system topology structure, and the affected user range and key load conditions are defined to obtain a fault impact assessment report, including: The fault section location result is spatially mapped with the distribution network topology data to determine the line section where the fault point is located and the electrical connection relationship, and obtain the network topology breakpoint location; Identify all power distribution nodes that cannot obtain power supply according to the network topology breakpoint location to obtain the precise power outage range boundary; Performing cluster analysis on the distribution transformer area data within the precise power outage range, stratifying them according to power supply capacity and load characteristics, and obtaining graded load impact data; Extracting the first-level security user information from the hierarchical load impact data, and combining it with the backup power supply configuration status to obtain an emergency power supply demand list; Calculating load loss quantitative indicators based on the precise blackout range boundary and the graded load impact data, establishing an economic loss model using a time weighting factor, and obtaining a time-division impact severity curve; The network topology breakpoint location, the precise power outage range boundary, the graded load impact data, the emergency power supply demand list and the time period impact severity curve are integrated into a hierarchical structure report to obtain a fault impact assessment report.

8. A traveling wave positioning system based on intelligent fuse, characterized in that: For implementing the traveling wave positioning method based on the smart fuse according to any one of claims 1 to 7, the traveling wave positioning system based on the smart fuse comprises: The sampling module is used to install intelligent fuses on the three phases of the distribution line and perform high-frequency sampling on the line signal to obtain digital data including power frequency signals and high-frequency signals; A calculation module, used for transmitting the digital data to a main control fuse, performing zero-sequence quantity synthesis calculation using the main control fuse, and obtaining zero-sequence voltage, zero-sequence current and fault judgment results; An extraction module, used for applying adaptive threshold wavelet transform processing to the collected line signal according to the zero-sequence voltage and the zero-sequence current, extracting multi-layer wavelet coefficients and constructing a feature vector matrix to obtain anti-interference enhanced traveling wave feature information; A distance measurement module is used to combine the zero-sequence current and the traveling wave characteristic information, and use an improved double-terminal positioning algorithm to perform distance measurement calculation on the fault point position to obtain a fault section positioning result; The grading module is used to execute a graded protection control strategy based on the fault judgment result and the fault section location result to obtain predictive maintenance information.

9. A traveling wave positioning device based on an intelligent fuse, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor implements the traveling wave positioning method based on the intelligent fuse according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the processor is enabled to execute the traveling wave positioning method based on the smart fuse according to any one of claims 1 to 7.

Images