Power distribution network single-phase grounding positioning method and device based on multi-source information fusion

Abstract

The invention relates to the technical field of power equipment safety operation and maintenance, in particular to a power distribution network single-phase grounding positioning method and device based on multi-source information fusion. Heterogeneous sensing nodes deployed in a layered and distributed mode are adopted to synchronously collect steady-state and transient-state electrical quantity data of a power distribution network, and preprocessing is carried out to generate a standardized data set; multi-dimensional feature extraction is carried out; carrying out weighted fusion on the extracted multi-dimensional features based on a multi-feature fusion strategy of a dynamic weight adjustment mechanism and a D-S evidence theory, and calculating a line fault probability; when the calculated line fault probability meets a preset threshold value, the position of a fault point is locked through first-stage line screening, second-stage section division and third-stage accurate positioning in sequence. Therefore, through a multi-stage data fusion and hierarchical positioning algorithm, the fault positioning error is obviously improved from the line level in the prior art to the section level, and the fault positioning problem under the scenes of high-resistance grounding, intermittent faults and a complex power distribution network containing a distributed power supply is effectively solved.

Description

Technical Field

[0001] This application relates to the field of power equipment safety operation and maintenance technology, and in particular to a method and device for locating single-phase grounding in power distribution networks based on multi-source information fusion. Background Technology

[0002] As a crucial component of the power system, the safe and stable operation of the distribution network directly impacts power supply reliability and power quality. Single-phase grounding faults are among the most common fault types in distribution network operation, accounting for over 70% of all distribution network faults. Current distribution network grounding fault handling technologies are mainly based on the following three working mechanisms: The first is the zero-sequence current detection method, which determines the fault location by monitoring the amplitude and phase changes of the zero-sequence current in the line. This method responds quickly in low-resistance grounding scenarios, but is greatly affected by the system's operating mode and is prone to misjudgment in complex topologies. The second is traveling wave localization technology, which utilizes the propagation characteristics of transient traveling waves generated by the fault to achieve location. Its location accuracy is theoretically high, but it has stringent requirements for the accuracy of the time synchronization system and is susceptible to interference from factors such as line branches and load changes. The third is multi-sensor fusion schemes, such as integrating multi-source data including current, voltage, and transient signals to improve location reliability. Its core lies in reducing the measurement error of a single sensor through data-layer fusion, but it does not involve deep fusion at the feature layer and decision layer.

[0003] In actual power distribution network operation, existing technologies have significant performance bottlenecks: First, they lack sensitivity in high-resistance grounding scenarios. When the grounding resistance exceeds 1000Ω, the zero-sequence current amplitude drops by more than 60% compared to metallic grounding, resulting in a failure rate of up to 35% for traditional overcurrent protection devices, making it impossible to effectively identify such faults. Second, the depth of multi-sensor data fusion is limited. Existing solutions mostly remain at the data splicing level. Differences in sensor sampling frequencies lead to data time synchronization errors exceeding 2ms, and the positioning accuracy can only reach the 500m level, which is insufficient to meet the needs of precise emergency repairs. Third, they have poor adaptability to complex topologies. In power distribution networks with distributed power sources, changes in the direction of fault current increase the failure probability of traditional direction-based positioning methods by 40%, making them unsuitable for the development trend of power distribution networks with high penetration rates of new energy sources. Summary of the Invention

[0004] To overcome the shortcomings of the existing technology, this application provides a method and device for locating single-phase grounding in distribution networks based on multi-source information fusion, which can realize accurate fault identification and location in complex scenarios, and improve the efficiency of fault handling and power supply reliability in distribution networks.

[0005] In a first aspect, this application provides a method for locating single-phase grounding in a distribution network based on multi-source information fusion, the method comprising the following steps: Heterogeneous sensing nodes deployed in a hierarchical and distributed manner synchronously collect steady-state and transient electrical quantity data of the power distribution network, and preprocess the collected electrical quantity data to generate a standardized dataset; Multi-dimensional features are extracted from the standardized dataset, and the extracted multi-dimensional features are weighted and fused based on a dynamic weight adjustment mechanism and a multi-feature fusion strategy of DS evidence theory to calculate the line fault probability. When the calculated line fault probability meets the preset threshold, the fault location is locked through first-level line screening, second-level segment division and third-level precise positioning.

[0006] In one possible implementation, the hierarchically distributed heterogeneous sensing nodes include: Transient waveform recording devices deployed on the substation side are used to collect high-frequency transient components; The intelligent fusion terminal, deployed in the middle section of the line, integrates zero-sequence voltage and current transformers and temperature and humidity sensors for real-time monitoring of the line's operating status. Deploy smart meters with zero-sequence current detection function on the user side to synchronously collect fundamental zero-sequence current and electricity consumption data.

[0007] In one possible implementation, the synchronous acquisition of steady-state and transient electrical quantity data of the distribution network, and the preprocessing of the acquired electrical quantity data to generate a standardized dataset, includes the following steps: For steady-state electrical quantity data of the distribution network, a periodic acquisition mechanism is adopted; for transient electrical quantity data of the distribution network, a fault-triggered acquisition mechanism is adopted. When a voltage change or zero-sequence current exceeds the limit is detected, high-speed waveform recording is automatically started to record the complete transient process before and after the fault for a set duration. A wavelet threshold denoising algorithm is used to denoise transient electrical quantity data, while effective value calculation and feature extraction are performed on steady-state electrical quantity data. The preprocessed electrical quantity data is uploaded to the cloud platform for time-scale alignment, data verification, and feature classification to generate a standardized dataset with a unified format. Among them, transient electrical quantity data is transmitted using the UDP protocol, while steady-state electrical quantity data is transmitted using the TCP protocol.

[0008] In one possible implementation, the dimensional features extracted from the standardized dataset include zero-order parameters, transient features, power features, and system state variables. The zero-sequence parameters include zero-sequence current amplitude, zero-sequence voltage phase angle, and zero-sequence impedance magnitude; the transient characteristics include transient energy entropy, transient current attenuation coefficient, and initial traveling wave front time; the power characteristics include active power abrupt change slope, reactive power fluctuation, and apparent power distortion rate; the system state quantities include neutral point displacement voltage, estimated grounding resistance, and pre-fault voltage fluctuation rate.

[0009] In one possible implementation, the multi-feature fusion strategy based on dynamic weight adjustment mechanism and DS evidence theory performs weighted fusion of extracted multi-dimensional features to calculate the line fault probability, including the following steps: A differentiated credibility evaluation function is constructed; among which, the credibility of zero-sequence current adopts a threshold normalization model, the credibility of transient characteristics adopts a signal stationarity measurement model, and the credibility of power characteristics adopts a dual threshold interval model. The credibility of each feature is mapped to a basic probability allocation function, and the Dempster combination rule is used for evidence synthesis. High-conflict features are corrected, and the weights of each piece of evidence are adaptively adjusted according to the current stage of the fault development to obtain the line fault probability. The fault development stage includes the initial stage, the transition stage, and the steady-state stage.

[0010] In one possible implementation, the zero-sequence current confidence level is:

[0011] in, The measured effective value of the zero-sequence current. Set a threshold for the system; The reliability of transient features is:

[0012] in, The 3σ variance of the transient signal; When the absolute value of the power mutation value ΔP is less than the minimum threshold When the power characteristic confidence α_P is 0, the absolute value of ΔP is between the minimum threshold. and When the power characteristic confidence α_P is between, Calculate; when the absolute value of ΔP is greater than At that time, the confidence level α_P of the power characteristic is 1; where the power mutation value is... , This is the power sampling value after the fault. This represents the average power value of the cycle preceding the fault.

[0013] In one possible implementation, the step of sequentially identifying the fault location through primary line screening, secondary segment division, and tertiary precise positioning includes the following steps: The primary line screening method calculates the fault probability of each line based on zero-sequence current and transient characteristics, using the following formula: ,in Let n be the failure probability value of the nth line; when When the value exceeds the set threshold, the line is locked as a candidate for fault. Secondary segment division, establishing a segment correlation matrix based on the distribution network topology. ,in This indicates that segments i and j have a direct electrical connection; multiple consecutive suspicious segments are identified based on the consistency of the fault current direction; where, if the fault current direction detection values ​​of adjacent segments i and j are respectively and ,and Indicates positive, This indicates the reverse direction, and the fault point is determined to be located between segments i and j; Three-level precise positioning, using an improved traveling wave positioning algorithm. Calculate the distance to the fault point, where To obtain the corrected traveling wave velocity, the theoretical wave velocity is first calculated based on the line type and parameters. The calculation is performed, where L is the inductance per unit length of the line and C is the capacitance per unit length of the line; then adaptive correction is performed based on ambient temperature and the service life of the line. The time difference between the arrival of the fault traveling wave front at the detection points at both ends of the line is extracted using the wavelet transform modulus maxima method; and the fault location is determined by combining the transformer area topology.

[0014] Secondly, this application provides a single-phase grounding location device for power distribution networks based on multi-source information fusion, the device comprising: The multi-dimensional data acquisition module is used to synchronously acquire steady-state and transient electrical quantity data of the power distribution network using heterogeneous sensing nodes deployed in a hierarchical and distributed manner, and to preprocess the acquired electrical quantity data to generate a standardized dataset. The fault probability calculation module is used to extract multi-dimensional features from the standardized dataset, and to perform weighted fusion of the extracted multi-dimensional features based on the dynamic weight adjustment mechanism and the DS evidence theory multi-feature fusion strategy to calculate the line fault probability. The fault location module is used to locate the fault point by sequentially performing first-level line screening, second-level segment division, and third-level precise positioning when the calculated line fault probability meets the preset threshold.

[0015] Thirdly, this application provides an electronic device, including: a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the bus, and when the machine-readable instructions are executed by the processor, the steps of the single-phase grounding location method for power distribution networks based on multi-source information fusion as described in any of the first aspects are performed.

[0016] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the single-phase grounding location method for power distribution networks based on multi-source information fusion as described in any of the first aspects.

[0017] This embodiment provides a method and device for single-phase grounding location in distribution networks based on multi-source information fusion. It employs heterogeneous sensing nodes deployed in a hierarchical, distributed manner to synchronously collect steady-state and transient electrical quantity data of the distribution network. The collected electrical quantity data is preprocessed to generate a standardized dataset. Multi-dimensional features are extracted from the standardized dataset, and a multi-feature fusion strategy based on a dynamic weight adjustment mechanism and DS evidence theory is used to weight and fuse the extracted multi-dimensional features to calculate the line fault probability. When the calculated line fault probability meets a preset threshold, the fault location is determined sequentially through first-level line screening, second-level section division, and third-level precise location. Thus, through multi-level data fusion and hierarchical location algorithms, the fault location error is significantly improved from the line level in existing technologies to the section level, effectively solving the fault location challenges in complex distribution network scenarios with high-resistance grounding, intermittent faults, and distributed power sources. This greatly improves the accuracy of fault location and provides decision support for rapid repair. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart of a single-phase grounding location method for a distribution network based on multi-source information fusion according to an embodiment of this application is shown; Figure 2 This document illustrates a flowchart of an embodiment of the present application, which describes a multi-feature fusion strategy based on a dynamic weight adjustment mechanism and DS evidence theory to perform weighted fusion of extracted multi-dimensional features and calculate the probability of line faults. Figure 3 This document illustrates a flowchart of an embodiment of the present application describing the process of sequentially identifying the fault location through primary line screening, secondary segment division, and tertiary precise positioning. Figure 4 This paper shows a schematic diagram of the structure of the single-phase grounding location device for power distribution networks based on multi-source information fusion as described in an embodiment of this application; Figure 5 A schematic diagram of the structure of the electronic device described in an embodiment of this application is shown. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0021] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0022] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0023] In view of the technical problems mentioned in the background, this application provides a method and device for single-phase grounding location in distribution networks based on multi-source information fusion, which can realize accurate fault identification and location in complex scenarios, and improve the efficiency of fault handling and power supply reliability in distribution networks.

[0024] In one embodiment, see the appendix to the specification. Figure 1 This application provides a method for locating single-phase grounding in a distribution network based on multi-source information fusion, comprising the following steps: S1. Use heterogeneous sensing nodes deployed in a hierarchical and distributed manner to synchronously collect steady-state and transient electrical quantity data of the distribution network, and preprocess the collected electrical quantity data to generate a standardized dataset. S2. Multi-dimensional features are extracted from the standardized dataset, and the extracted multi-dimensional features are weighted and fused based on a dynamic weight adjustment mechanism and a multi-feature fusion strategy using DS evidence theory to calculate the line fault probability. This step aims to extract key information characterizing grounding faults from massive amounts of data and form a comprehensive decision through information fusion technology. The specific process is as follows: S3. When the calculated line fault probability meets the preset threshold, the fault location is locked by sequentially going through first-level line screening, second-level section division and third-level precise positioning.

[0025] Specifically, in step S1, steady-state and transient electrical quantity data are captured through layered deployment of sensing nodes; and the collected raw data is transformed into a standardized dataset in a unified format to provide high-quality input for subsequent feature extraction.

[0026] In one embodiment, based on the distribution network topology (main line, branch line, load concentration area), the number and installation location of data acquisition nodes on the substation side, the middle section of the line, and the user side are determined according to the principle of "closer and more dense fault points and more sparse fault points".

[0027] On the substation side: Deploy a high-precision transient waveform recording device with a 1MHz sampling rate and 24-bit AD conversion accuracy to continuously monitor the electrical quantities of the line, focusing on capturing the high-frequency transient components of 2-10kHz at the moment of a fault, and simultaneously recording the acquisition timestamp.

[0028] Mid-section of the line: Install intelligent fusion terminals. Through the zero-sequence voltage / current transformers of the intelligent fusion terminals, collect steady-state zero-sequence voltage and zero-sequence current data of the line in real time. Temperature and humidity sensors record environmental parameters simultaneously. The terminal is in a low-power monitoring state to monitor voltage surges and zero-sequence current over-limit signals.

[0029] On the user side: Deploy smart meters with zero-sequence current detection function to collect 50 Hz fundamental zero-sequence current while measuring electricity consumption data, forming a peripheral sensing network for fault monitoring.

[0030] Based on the time-frequency domain distribution characteristics of the fault features, the collected data are divided into two core types, as shown in Table 1.

[0031]

[0032] Table 1 This approach utilizes a layered deployment of three types of data acquisition nodes to cover the entire power distribution network from source to end: substation-side devices focus on high-frequency transient components, providing raw data support for fault location; mid-line terminals enable real-time monitoring of operational status, compensating for line blind spots; and user-side smart meters form a terminal sensing network, increasing the density of fault signal capture. Transient data is acquired using a "fault-triggered" acquisition method, ensuring complete recording of the fault process while reducing equipment power consumption and data redundancy in non-fault states. The amount of data acquired is controlled at 200KB per acquisition, balancing data integrity and transmission efficiency.

[0033] Furthermore, the collected electrical quantity data is subjected to denoising, compression, transmission and standardization processing to transform the messy raw data into a standardized dataset with a unified format, providing high-quality input for subsequent feature extraction and solving the problems of inconsistent formats, interference and excessive size of multi-source data.

[0034] In one embodiment, to address the bandwidth bottleneck and main station computing power pressure of massive heterogeneous data transmission, a two-tier processing architecture of edge preprocessing and cloud-based deep analysis is constructed. For example, through the built-in edge computing module, a wavelet threshold denoising algorithm is used to preprocess transient electrical quantity data, removing white noise and interference signals, achieving a data compression ratio of up to 10:1. Simultaneously, effective value calculation and feature extraction are performed on steady-state electrical quantity data to reduce the original data volume. The preprocessed transient and steady-state electrical quantity data are uploaded to the cloud platform via a 5G / fiber hybrid network. The transient data uses the UDP protocol for high-speed transmission, while the steady-state data uses the TCP protocol to ensure transmission reliability, with transmission latency controlled within 200 ms. After receiving data from each node, the cloud platform performs fusion processing according to time-stamp alignment, data verification, and feature classification, providing a standardized dataset for subsequent fault diagnosis algorithms.

[0035] The main station platform extracts the local timestamps of all data and corrects them based on the substation-side GPS synchronization clock (accuracy ≤1μs) and the historical synchronization errors of each node. For missing data periods, linear interpolation is used to complete the data, ensuring timeline continuity. Then, integrity checks (verifying data packet completeness and triggering retransmission at edge nodes if missing data), validity checks (removing outliers outside reasonable ranges), and redundancy checks (correcting sensor fault data through cross-validation of data from adjacent nodes) are performed. Finally, the verified data is divided into steady-state and transient datasets, and the data units are standardized, converted to floating-point format, and a standardized dataset is generated.

[0036] In step S2, the main focus is on extracting effective fault features from the data, achieving a deeper understanding from the data layer to the feature layer. This application extracts 12 core features covering zero-sequence parameters, transient characteristics, power characteristics, and system state quantities to comprehensively characterize the electrical properties of the fault, avoiding the limitations of single features. The zero-sequence parameters include zero-sequence current amplitude, zero-sequence voltage phase angle, and zero-sequence impedance magnitude; the transient characteristics include transient energy entropy, transient current attenuation coefficient, and initial traveling wave front time; the power characteristics include active power abrupt change slope, reactive power fluctuation, and apparent power distortion rate; and the system state quantities include neutral point displacement voltage, estimated grounding resistance, and pre-fault voltage fluctuation rate.

[0037] Specifically, the calculation formulas and physical meanings of the 12 core feature quantities covering zero-sequence parameters, transient characteristics, power characteristics, and system state quantities are as follows: Zero-sequence parameter: zero-sequence current amplitude (Three-phase current vector sum).

[0038] Zero-sequence voltage phase angle ;in The zero-sequence voltage phase angle, The phase angle of phase A voltage.

[0039] Zero-sequence impedance magnitude ; Transient characteristics: Transient energy entropy ,in This represents the energy percentage of each frequency band after wavelet packet decomposition.

[0040] Transient current attenuation coefficient By analyzing the current curve after the fault The results were obtained through fitting.

[0041] Initial traveling wave head time The maximum point of the modulus is detected by db4 wavelet transform.

[0042] Power characteristics: Active power change slope The difference in active power of the frequency before and after the fault. This represents the time interval between power fluctuations.

[0043] reactive power fluctuation ;in For the first The reactive power value at each sampling point The reactive power value is set within a specified time period before and after the fault, and N is the total number of sampling points.

[0044] Apparent power distortion ;in For the first Apparent power of subharmonics, Here, H represents the apparent power of the fundamental frequency, and H represents the highest harmonic order.

[0045] System state variables: Neutral point displacement voltage (Direct measurement).

[0046] Estimated grounding resistance .

[0047] Voltage fluctuation before fault (Standard deviation of the effective value of the voltage 10 cycles before the fault).

[0048] The above 12 characteristic values ​​are then transformed into a unified line fault probability. In one embodiment, see the appendix to the specification. Figure 2 The multi-feature fusion strategy based on dynamic weight adjustment mechanism and DS evidence theory performs weighted fusion of extracted multi-dimensional features to calculate the line fault probability, including the following steps: S201. Construct a differentiated credibility evaluation function; wherein, the zero-sequence current credibility adopts a threshold normalization model, the transient characteristic credibility adopts a signal stationarity measurement model, and the power characteristic credibility adopts a dual threshold interval model. S202. Map the credibility of each feature to a basic probability allocation function and use the Dempster combination rule to synthesize evidence; correct high-conflict features and adaptively adjust the weights of each piece of evidence according to the current stage of the fault development to obtain the line fault probability; wherein, the fault development stage includes the initial stage, the transition stage, and the steady-state stage.

[0049] Specifically, steps S201 and S202 include three processes: feature credibility assessment, DS evidence theory fusion, and dynamic weight adjustment.

[0050] (a) Feature credibility assessment: Using a differentiation model to divide each feature value Mapping to credibility .

[0051] Zero-sequence current confidence level (threshold normalization model): , where the threshold Take 15% of the line's rated current.

[0052] Transient feature reliability (signal stationarity measurement model): 3σ is the standard deviation of the high-frequency components of the transient signal after wavelet denoising.

[0053] Power feature reliability (dual threshold interval model): definition , This represents the average active power over the previous week before the fault.

[0054]

[0055] (b) Integration of DS Evidence Theory: Constructing the identification framework and basic probability assignment: Define the identification framework Θ = {F, N}, where F represents "single-phase ground fault occurred" and N represents "no fault occurred". For each feature i, its basic probability assignment (mass) function is:

[0056] in, For the first The basic probability assignment function for each feature For the first The credibility of each feature.

[0057] Evidence synthesis and conflict detection: Dempster's combination rule is used to sequentially synthesize all 12 mass functions. For any two pieces of evidence... and Its synthesis formula is: .

[0058] in, The degree of confidence in propositions B and C based on two independent pieces of evidence, and the coefficient of conflict. Quantify the degree of contradiction between pieces of evidence.

[0059] High-conflict evidence correction: When When this condition is met, it is considered a high-conflict condition. A discount factor method is used for correction. The revised evidence is as follows: .

[0060] (c) Dynamic weight adjustment: Based on the stage of fault development, the above fusion results are adjusted twice in terms of weight to highlight the role of the dominant characteristics at different stages: Initial stage (t<0.04s): The total weight of transient features increases to 80%.

[0061] Transition phase (0.04s ≤ t < 0.1s): Transient and steady-state characteristics each account for 40% of the weight.

[0062] Steady-state phase (t ≥ 0.1s): The total weight of steady-state characteristics (zero-sequence parameters, system state variables) is increased to 60%.

[0063] Weight adjustment is achieved using linear interpolation: ;in, for Time of the first Dynamic weights of each feature For the first The initial weights of each feature, For the first The steady-state weights of each feature These are the linear interpolation coefficients. The adjusted weights are combined with the corrected evidence, and Dempster synthesis is performed again to obtain the final fusion probability allocation. The probability of a line fault is: .

[0064] In step S3, a hierarchical and progressive fault location strategy is mainly adopted, which achieves accurate location of the fault point through multi-dimensional electrical quantity fusion and logical reasoning.

[0065] In one embodiment, see the appendix to the specification. Figure 3 The process of sequentially identifying the fault location through primary line screening, secondary section division, and tertiary precise positioning includes the following steps: S301. Primary line screening: The fault probability of each line is calculated based on zero-sequence current and transient characteristics. The formula is as follows: ,in Let n be the failure probability value of the nth line; when When the value exceeds the set threshold, the line is locked as a candidate for fault. S302. Second-level segment division, establishing segment correlation matrix. ,in, This indicates that there is a direct electrical connection between sections i and j; multiple consecutive suspicious sections are identified based on the consistency of the fault current direction. S303, Level 3 precise positioning, employing an improved traveling wave positioning algorithm. ,in, The corrected traveling wave velocity, The distance to the fault point is calculated based on the arrival time difference of the traveling wave front of the fault, and the fault location is locked by combining the topology of the transformer area.

[0066] Specifically, when a certain line (This threshold is determined based on historical data and corresponds to an accuracy rate of over 95%), when the threshold is reached, the line is determined to be a candidate line for fault, triggering the three-level location process. Specifically, in step S301, during the first-level line screening, on the substation outgoing line side, based on two core features—zero-sequence current amplitude and transient energy entropy—the fault probability of each outgoing line is quickly calculated. The highest probability value is selected as the risk indicator for this line group. .Will The highest-ranking line is identified as a potential fault line, completing the initial location from "area" to "line".

[0067] In step S302, during the secondary segment division, the fault current direction detection unit installed at the sectionalizing switch is used to locate the segment on the identified faulty line. A segment association matrix is ​​generated based on the distribution network topology GIS data. If the section and If directly connected, then Otherwise, it is 0. Read the direction of the fault current detected by each switch. (Define the flow direction to the end of the line as positive +1, and the flow direction to the bus as negative -1). Perform consistency reasoning: if adjacent segments and satisfy and If so, the fault point is determined to be located in the section. and Based on this, the fault range can be narrowed down to 2-3 consecutive sections.

[0068] In step S303, when performing three-level precise positioning, an improved double-ended traveling wave positioning method is used to achieve meter-level precision positioning in the suspicious section.

[0069] Traveling wave front detection and synchronization: Collect transient voltage data from monitoring points at both ends of the section, and use the db4 wavelet transform modulus maxima method to accurately extract the arrival time of the initial traveling wave of the fault. and The two devices synchronize their time via GPS / BDS, with an error... Calculate the time difference .

[0070] Traveling wave velocity adaptive correction: theoretical wave velocity Real-time correction is performed considering the actual ambient temperature T (°C) and the line's service life Y (years):

[0071] Distance calculation and location locking: and the corrected Substitute into the traveling wave positioning formula: The precise distance d from the fault point to one end of the measurement point is calculated. Combined with the precise Geographic Information System (GIS) coordinates of the towers and branch nodes within this section, the fault point can be pinpointed to a specific tower or branch line, with the location error controlled within [specific range]. Within meters.

[0072] This application provides a method for locating single-phase ground faults in distribution networks based on multi-source information fusion. Employing multi-source data fusion technology and integrating electrical quantity monitoring data, it achieves a 25 percentage point improvement over traditional single-feature detection methods, effectively solving the identification challenges of difficult scenarios such as high-resistance grounding and intermittent faults. Furthermore, the innovatively designed three-level location model (regional location - section location - precise point location) achieves a step-by-step reduction of the fault range, ultimately controlling the investigation area within 500m. In a 10kV distribution network simulation platform verification, the location error in a high-resistance grounding (1000 Ω) scenario is ≤300 m, and the fault isolation time is shortened to 15 minutes, representing an 8-16 times increase in efficiency compared to the traditional manual line inspection mode (2-4 hours), significantly reducing the duration of power outages.

[0073] In the implementation scenario of a 10 kV neutral point grounded system via an arc suppression coil: (1) Configuration of core hardware nodes.

[0074] Substation side: A transient waveform recorder is installed on the 10kV side of the main transformer. The sampling rate is set to 1MHz, the AD accuracy is 24-bit, and it has 64 analog input channels. It continuously monitors the electrical quantities of each outgoing line, focuses on capturing high-frequency transient components of 2-10kHz, and supports data recording of 5 cycles before and after a fault.

[0075] Mid-section of the line: Intelligent integrated terminals are deployed at 15 sectional switches on the main line. Each terminal integrates a zero-sequence voltage / current transformer (100 / 5 ratio, 0.5 accuracy) and a temperature and humidity sensor to collect the zero-sequence electrical quantities of the line and the ambient temperature and humidity status in real time.

[0076] On the user side: Smart meters conforming to the DL / T 645-2007 protocol (with zero-sequence current detection function and measurement error ≤1%) are installed at 128 user access points at the end of the distribution network to synchronously collect 50Hz fundamental zero-sequence current and user electricity consumption data, forming a terminal sensing network.

[0077] Each node achieves time synchronization through a built-in GPS / BDS module with a synchronization error of ≤1μs, and connects to the cloud main station platform through a 5G / fiber hybrid network.

[0078] (2) Data acquisition and preprocessing.

[0079] The system employs a mechanism that combines periodic data acquisition with fault-triggered data acquisition. Steady-state data: The effective values ​​of three-phase voltage / current, power, and zero-sequence component are collected at 1-second intervals.

[0080] Transient data: When a voltage surge exceeding 15% or a zero-sequence current exceeding a set threshold (15% of the line's rated current, i.e., 5A) is detected, high-speed waveform recording is automatically triggered at all nodes along the line to record the complete transient process for 10 cycles before and after the fault.

[0081] On the edge side, a wavelet thresholding denoising algorithm (db4 wavelet basis, 5-level decomposition) is used to preprocess transient electrical quantity data to remove white noise and power frequency interference, achieving a data compression ratio of 10:1. Simultaneously, effective values ​​are calculated and preliminary features are extracted from steady-state electrical quantity data. The preprocessed data is uploaded to the cloud platform: transient data uses UDP protocol for high-speed, low-latency transmission; steady-state data uses TCP protocol to ensure reliable transmission. After receiving the data, the cloud platform performs time-stamp alignment, data verification (integrity, validity, and redundancy checks), and feature classification to generate a standardized dataset with a unified format.

[0082] (3) Multi-dimensional feature extraction and fusion calculation.

[0083] Taking a real single-phase ground fault as an example, the system extracts 12 feature quantities from the standardized dataset, with the core values ​​as follows: Zero-sequence parameter: zero-sequence current amplitude Zero-sequence voltage phase angle Zero-sequence impedance magnitude .

[0084] Transient characteristics: Transient energy entropy transient current attenuation coefficient Initial traveling wave head time The fault was detected 153.250ms after the db4 wavelet transform.

[0085] Power characteristics: slope of active power abrupt change reactive power fluctuation Apparent power distortion .

[0086] System state variables: Neutral point displacement voltage Estimated grounding resistance Voltage fluctuation rate before the fault .

[0087] The probability of line faults is calculated by multi-feature fusion based on a dynamic weight adjustment mechanism and DS evidence theory. 1. Differentiated Credibility Assessment: Zero-sequence current reliability: .

[0088] Reliability of transient features: Calculate the standard deviation of high-frequency components of the transient signal. ,but .

[0089] Power characteristic reliability: power mutation value .set up ,because Therefore .

[0090] 2. DS Evidence Synthesis and Dynamic Weight Adjustment: Each feature's credibility is mapped to a basic probability assignment function, and Dempster's combination rule is used for sequential evidence synthesis. The time of failure is the initial stage. Based on a dynamic weight adjustment strategy, the total weight of transient feature-type evidence is increased to 80%. After synthesis and adaptive weight adjustment, the final fused failure probability of the L2 eligibility level is obtained. It exceeds the preset threshold of 0.85.

[0091] (4) Three-level progressive positioning process.

[0092] The system automatically triggers the three-level positioning process: Level 1 line screening: Calculate the failure probability of each outgoing line to obtain... This corresponds to line L2. L2 is determined to be a candidate fault line, completing the initial location from the surface to the line.

[0093] Secondary segment division: Generate segment correlation matrix based on distribution network topology. Read the fault current direction detected by each intelligent fusion terminal on the L2 line: Detection value of section S5 (near the busbar side). (Flowing towards the end of the line), detection value of adjacent section S6 (far busbar side) (Flowing towards the busbar). Based on the directional consistency criterion ( The fault point was determined to be located between sections S5 and S6, narrowing the fault range to two consecutive sections (approximately 1.2 kilometers).

[0094] Level 3 Precise Location: Voltage transient data were collected at monitoring points at both ends of the suspected section (S5-S6). The arrival time of the initial traveling wavefront of the fault was accurately extracted using the db4 wavelet transform modulus maxima method. , Time difference .

[0095] Theoretical wave speed of the line Based on real-time monitored ambient temperature and line service life Adaptive correction is performed annually:

[0096] Substitute the values ​​into the improved traveling wave location formula to calculate the distance to the fault point:

[0097] Based on the precise GIS coordinates of the towers in this section, the fault point was determined to be located between tower #86 (3000 meters from monitoring point S5) and tower #87 (3200 meters from monitoring point S5), approximately 157 meters from tower #86.

[0098] Based on the same inventive concept, this application also provides a distribution network single-phase grounding location device based on multi-source information fusion. Since the principle of the device in this application is similar to the above-mentioned distribution network single-phase grounding location method based on multi-source information fusion in this application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0099] As per the instruction manual Figure 4 As shown in the figure, this application embodiment also provides a single-phase grounding location device for power distribution networks based on multi-source information fusion, the device comprising: The multi-dimensional data acquisition module 401 is used to synchronously acquire steady-state and transient electrical quantity data of the power distribution network using heterogeneous sensing nodes deployed in a hierarchical and distributed manner, and to preprocess the acquired electrical quantity data to generate a standardized dataset. The fault probability calculation module 402 is used to extract multi-dimensional features from the standardized dataset, and to perform weighted fusion of the extracted multi-dimensional features based on the dynamic weight adjustment mechanism and the DS evidence theory multi-feature fusion strategy to calculate the line fault probability. The fault point locking module 403 is used to lock the location of the fault point by sequentially performing first-level line screening, second-level segment division and third-level precise positioning when the calculated line fault probability meets the preset threshold.

[0100] In some embodiments, the hierarchically distributed heterogeneous sensing nodes include: a transient waveform recording device deployed on the substation side for collecting high-frequency transient components; an intelligent fusion terminal deployed in the middle section of the line, integrating a zero-sequence voltage and current transformer and a temperature and humidity sensor for real-time monitoring of the line's operating status; and a smart meter with zero-sequence current detection function deployed on the user side for synchronously collecting fundamental zero-sequence current and electricity consumption data.

[0101] In some embodiments, the multi-dimensional data acquisition module 401 synchronously acquires steady-state and transient electrical quantity data of the distribution network, and preprocesses the acquired electrical quantity data to generate a standardized dataset, including: employing a periodic acquisition mechanism for steady-state electrical quantity data of the distribution network; employing a fault-triggered acquisition mechanism for transient electrical quantity data of the distribution network, automatically starting high-speed waveform recording when a voltage surge or zero-sequence current exceedance is detected, recording the complete transient process before and after the fault for a set duration; employing a wavelet threshold denoising algorithm to denoise the transient electrical quantity data, while simultaneously calculating the effective value and extracting features from the steady-state electrical quantity data; uploading the preprocessed electrical quantity data to a cloud platform for time-stamp alignment, data verification, and feature classification to generate a standardized dataset with a unified format; wherein, the transient electrical quantity data is transmitted using the UDP protocol, and the steady-state electrical quantity data is transmitted using the TCP protocol.

[0102] In some embodiments, the dimensional features extracted from the standardized dataset include zero-sequence parameters, transient features, power features, and system state variables; the zero-sequence parameters include zero-sequence current amplitude, zero-sequence voltage phase angle, and zero-sequence impedance magnitude; the transient features include transient energy entropy, transient current decay coefficient, and initial traveling wave front time; the power features include active power abrupt change slope, reactive power fluctuation, and apparent power distortion rate; the system state variables include neutral point displacement voltage, estimated grounding resistance, and pre-fault voltage fluctuation rate.

[0103] In some embodiments, the fault probability calculation module 402 performs weighted fusion of extracted multi-dimensional features based on a dynamic weight adjustment mechanism and a multi-feature fusion strategy of DS evidence theory to calculate the line fault probability, including: constructing a differentiated credibility evaluation function; wherein, the zero-sequence current credibility adopts a threshold normalization model, the transient feature credibility adopts a signal stationarity measurement model, and the power feature credibility adopts a dual threshold interval model; mapping each feature credibility to a mass function, correcting high-conflict features, and adaptively adjusting the weights of the corrected feature credibility according to the current fault development stage to obtain the line fault probability; wherein, the fault development stage includes the fault initial stage, the transition stage, and the steady-state stage.

[0104] The single-phase grounding location device for power distribution networks based on multi-source information fusion described in this application synchronously collects steady-state and transient electrical quantity data of the power distribution network through a multi-dimensional data acquisition module using heterogeneous sensing nodes deployed in a hierarchical distributed manner. The collected electrical quantity data is preprocessed to generate a standardized dataset. A fault probability calculation module extracts multi-dimensional features from the standardized dataset and performs weighted fusion of the extracted multi-dimensional features based on a dynamic weight adjustment mechanism and DS evidence theory to calculate the line fault probability. When the calculated line fault probability meets a preset threshold, the fault location is located sequentially through primary line screening, secondary section division, and tertiary precise positioning. Thus, through multi-level data fusion and hierarchical positioning algorithms, the fault location error is significantly improved from the line level of existing technologies to the section level, greatly enhancing the accuracy of fault location and providing decision support for rapid repair.

[0105] Based on the same concept of the present invention, as shown in the appendix to the specification. Figure 5 As shown in the embodiment of this application, an electronic device 500 is provided. The electronic device 500 includes: at least one processor 501, at least one network interface 504 or other user interface 503, a memory 505, and at least one communication bus 502. The communication bus 502 is used to enable communication between these components. The electronic device 500 may optionally include a user interface 503, including a display (e.g., touchscreen, LCD, CRT, holographic imaging, or projector), a keyboard, or a clicking device (e.g., mouse, trackball, touchpad, or touchscreen).

[0106] Memory 505 may include read-only memory and random access memory, and provides instructions and data to processor 501. A portion of memory 505 may also include non-volatile random access memory (NVRAM).

[0107] In some implementations, memory 505 stores executable modules or data structures, or subsets thereof, or extended sets thereof: The 5051 operating system contains various system programs used to implement various basic business functions and handle hardware-based tasks. Application module 5052 contains various applications, such as launchers, media players, and browsers, to implement various application functions.

[0108] In this embodiment of the application, by calling the program or instructions stored in the memory 505, the processor 501 is used to execute steps such as a method for locating single-phase grounding in a power distribution network based on multi-source information fusion.

[0109] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs steps such as those in a single-phase grounding location method for a power distribution network based on multi-source information fusion.

[0110] Specifically, the storage medium can be a general-purpose storage medium, such as a portable disk or hard drive. When the computer program on the storage medium is run, it can realize the entire process from system diagram drawing to bill of materials creation, thereby optimizing supply chain management.

[0111] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, and the indirect coupling or communication connection of the apparatus or units may be electrical, mechanical, or other forms.

[0112] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0113] In addition, the functional units in the embodiments provided in this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0114] If a function is implemented as 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 this application, in essence, or the part that contributes to the prior art, or a part of the 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 cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0115] Finally, it should be noted that the above embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A method for locating single-phase ground faults in a distribution network based on multi-source information fusion, characterized in that, The method includes the following steps: Heterogeneous sensing nodes deployed in a hierarchical and distributed manner synchronously collect steady-state and transient electrical quantity data of the power distribution network, and preprocess the collected electrical quantity data to generate a standardized dataset; Multi-dimensional features are extracted from the standardized dataset, and the extracted multi-dimensional features are weighted and fused based on a dynamic weight adjustment mechanism and a multi-feature fusion strategy of DS evidence theory to calculate the line fault probability. When the calculated line fault probability meets the preset threshold, the fault location is locked through first-level line screening, second-level segment division and third-level precise positioning.

2. The method for locating single-phase grounding in a distribution network based on multi-source information fusion according to claim 1, characterized in that, in, The heterogeneous sensing nodes deployed in a hierarchical distributed manner include: Transient waveform recording devices deployed on the substation side are used to collect high-frequency transient components; The intelligent fusion terminal, deployed in the middle section of the line, integrates zero-sequence voltage and current transformers and temperature and humidity sensors for real-time monitoring of the line's operating status. Deploy smart meters with zero-sequence current detection function on the user side to synchronously collect fundamental zero-sequence current and electricity consumption data.

3. The method for locating single-phase grounding in a distribution network based on multi-source information fusion according to claim 1, characterized in that, The synchronous acquisition of steady-state and transient electrical quantity data of the distribution network, and the preprocessing of the acquired electrical quantity data to generate a standardized dataset, includes the following steps: For steady-state electrical quantity data of the distribution network, a periodic acquisition mechanism is adopted; for transient electrical quantity data of the distribution network, a fault-triggered acquisition mechanism is adopted. When a voltage change or zero-sequence current exceeds the limit is detected, high-speed waveform recording is automatically started to record the complete transient process before and after the fault for a set duration. A wavelet threshold denoising algorithm is used to denoise transient electrical quantity data, while effective value calculation and feature extraction are performed on steady-state electrical quantity data. The preprocessed electrical quantity data is uploaded to the cloud platform for time-scale alignment, data verification, and feature classification to generate a standardized dataset with a unified format. Among them, transient electrical quantity data is transmitted using the UDP protocol, while steady-state electrical quantity data is transmitted using the TCP protocol.

4. The method for locating single-phase grounding in a distribution network based on multi-source information fusion according to claim 1, characterized in that, in, The dimensional features extracted from the standardized dataset include zero-order parameters, transient features, power features, and system state variables; The zero-sequence parameters include zero-sequence current amplitude, zero-sequence voltage phase angle, and zero-sequence impedance magnitude; the transient characteristics include transient energy entropy, transient current attenuation coefficient, and initial traveling wave front time; the power characteristics include active power abrupt change slope, reactive power fluctuation, and apparent power distortion rate; the system state quantities include neutral point displacement voltage, estimated grounding resistance, and pre-fault voltage fluctuation rate.

5. The method for locating single-phase grounding in a distribution network based on multi-source information fusion according to claim 4, characterized in that, The multi-feature fusion strategy based on dynamic weight adjustment mechanism and DS evidence theory performs weighted fusion of extracted multi-dimensional features to calculate the line fault probability, including the following steps: A differentiated credibility evaluation function is constructed; among which, the credibility of zero-sequence current adopts a threshold normalization model, the credibility of transient characteristics adopts a signal stationarity measurement model, and the credibility of power characteristics adopts a dual threshold interval model. The credibility of each feature is mapped to a basic probability allocation function, and the Dempster combination rule is used for evidence synthesis. High-conflict features are corrected, and the weights of each piece of evidence are adaptively adjusted according to the current stage of the fault development to obtain the line fault probability. The fault development stage includes the initial stage, the transition stage, and the steady-state stage.

6. The method for locating single-phase grounding in a distribution network based on multi-source information fusion according to claim 5, characterized in that, The reliability of zero-sequence current is: in, The measured effective value of the zero-sequence current. Set a threshold for the system; The reliability of transient features is: in, The 3σ variance of the transient signal; When the absolute value of the power mutation value ΔP is less than the minimum threshold When the power characteristic confidence α_P is 0, the absolute value of ΔP is between the minimum threshold. and When the power characteristic confidence α_P is between, Calculate; when the absolute value of ΔP is greater than At that time, the confidence level α_P of the power characteristic is 1; where the power mutation value is... , This is the power sampling value after the fault. This represents the average power value of the cycle preceding the fault.

7. The method for locating single-phase grounding in a distribution network based on multi-source information fusion according to claim 6, characterized in that, The process of sequentially identifying the fault location through primary line screening, secondary section division, and tertiary precise positioning includes the following steps: The primary line screening method calculates the fault probability of each line based on zero-sequence current and transient characteristics, using the following formula: ,in Let n be the failure probability value of the nth line; when When the value exceeds the set threshold, the line is locked as a candidate for fault. Secondary segment division, establishing a segment correlation matrix based on the distribution network topology. ,in This indicates that segments i and j have a direct electrical connection; multiple consecutive suspicious segments are identified based on the consistency of the fault current direction; where, if the fault current direction detection values ​​of adjacent segments i and j are respectively and ,and Indicates positive, This indicates the reverse direction, and the fault point is determined to be located between segments i and j; Three-level precise positioning, using an improved traveling wave positioning algorithm. Calculate the distance to the fault point, where To obtain the corrected traveling wave velocity, the theoretical wave velocity is first calculated based on the line type and parameters. The calculation is performed, where L is the inductance per unit length of the line and C is the capacitance per unit length of the line; then adaptive correction is performed based on ambient temperature and the service life of the line. The time difference between the arrival of the fault traveling wave front at the detection points at both ends of the line is extracted using the wavelet transform modulus maxima method; and the fault location is determined by combining the transformer area topology.

8. A single-phase grounding location device for a power distribution network based on multi-source information fusion, characterized in that, The device includes: The multi-dimensional data acquisition module is used to synchronously acquire steady-state and transient electrical quantity data of the power distribution network using heterogeneous sensing nodes deployed in a hierarchical and distributed manner, and to preprocess the acquired electrical quantity data to generate a standardized dataset. The fault probability calculation module is used to extract multi-dimensional features from the standardized dataset, and to perform weighted fusion of the extracted multi-dimensional features based on the dynamic weight adjustment mechanism and the DS evidence theory multi-feature fusion strategy to calculate the line fault probability. The fault location module is used to locate the fault point by sequentially performing first-level line screening, second-level segment division, and third-level precise positioning when the calculated line fault probability meets the preset threshold.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, they perform the steps of the single-phase grounding location method for power distribution networks based on multi-source information fusion as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the single-phase grounding location method for distribution networks based on multi-source information fusion as described in any one of claims 1 to 7.

Images