Power transmission line fault detection and positioning method and system based on sensor network

Abstract

The invention is suitable for the field of automatic monitoring of a power system, and provides a power transmission line fault detection and positioning method and system based on a sensor network, and the method comprises the steps: continuously collecting multi-dimensional monitoring data through a sensor; each sensor independently judges abnormity, marks abnormal nodes and generates an abnormal information packet; the abnormal node broadcasts information, and the neighbor node verifies and generates verification information containing coordinates, voting types and confidence weights; the verification information is integrated, and a fault consensus is achieved if the sum of confidence weights exceeds a dynamic threshold value; and if consensus is reached, calculating and correcting the fault position, and if not, discarding data and processing false alarms. The system comprises corresponding acquisition, judgment, broadcasting, consensus judgment and positioning processing modules. The method improves the fault detection reliability and positioning precision, reduces the false alarm, and is suitable for a complex environment.

Description

Technical Field

[0001] This invention belongs to the field of power system automation monitoring, and particularly relates to a method and system for transmission line fault detection and location based on sensor networks. Background Technology

[0002] Transmission lines are the core carriers for long-distance power transmission in power systems, and their operational status directly determines the continuity and reliability of power supply. As power networks expand towards high voltage, long distances, and coverage of complex terrain, the risks faced by transmission lines, such as natural interference, mechanical losses, and external damage, have significantly increased. Traditional fault management models relying on manual inspections and fixed-point monitoring are no longer sufficient to meet the power system's demands for real-time fault detection and accurate fault location. In recent years, sensor network technology, with its advantages of distributed deployment, multi-node collaboration, and real-time data interaction, has gradually become a core technology direction for transmission line monitoring. The industry is widely promoting the integration of sensor networks and power monitoring, hoping to achieve rapid fault identification and accurate location through multi-dimensional data collection and analysis, thereby improving the operational efficiency and power supply stability of the power system.

[0003] Existing technical solutions suffer from insufficient reliability in fault detection, limited positioning accuracy, and poor environmental adaptability. Specifically, the data monitoring dimensions are limited, failing to fully capture the differentiated characteristics of electrical and mechanical faults, which can easily lead to missed fault detections. Furthermore, the lack of multi-node collaborative verification and dynamic consensus mechanisms means that misjudgments by a single node or momentary interference can easily trigger false alarms. Fixed fault judgment thresholds cannot adapt to changes in operating environments across different regions and seasons. Fault positioning does not incorporate data reliability weights and fault type characteristics, resulting in significant positioning errors. Moreover, the absence of effective false alarm suppression and self-optimization mechanisms leads to a decline in system performance over long-term operation, making it difficult to meet the fault management needs of transmission lines in complex environments. Summary of the Invention

[0004] The purpose of this invention is to provide a method for detecting and locating faults in power transmission lines based on sensor networks, aiming to solve the technical problems identified in the prior art in the background section.

[0005] This invention is implemented as follows: a method for detecting and locating faults in power transmission lines based on sensor networks, the method comprising: The transmission line continuously collects multi-dimensional monitoring data through sensors, including electrical quantity data and physical quantity data. Based on multi-dimensional monitoring data, each sensor independently and in real time determines whether a line abnormality has occurred. When any sensor detects a line abnormality event, it marks the sensor node as an abnormal node and generates an abnormal information packet containing a node identifier, abnormality type, and abnormality intensity value. An abnormal node is used as the initiating node to broadcast an abnormal information packet to all neighboring nodes within the communication range. The neighboring nodes that receive the abnormal information packet act as verification nodes, and perform local verification of the abnormal information packet based on their own collected local multi-dimensional monitoring data. They calculate the correlation between the multi-dimensional monitoring data and the abnormal information packet, and generate verification information including node coordinates, voting type and confidence weight. Centered on the initiating node, the verification information of all verification nodes within a preset range is integrated. Based on the confidence weight in the verification information, when the sum of the confidence weights exceeds the dynamic consensus threshold, a fault consensus is determined to be reached. When a fault consensus is reached, the fault location coordinates are calculated based on the node coordinates and confidence weights of all verification nodes; when a fault consensus is not reached, all relevant data of the line abnormal event are automatically discarded, and false alarm suppression processing is performed.

[0006] As a further embodiment of the present invention, the continuous acquisition of multi-dimensional monitoring data of the transmission line by sensors includes: The system continuously collects current and voltage waveform data, vibration acceleration data, and acoustic signal data of the transmission line through sensors, serving as multi-dimensional monitoring data. The collected multi-dimensional monitoring data is preprocessed, and the preprocessed multi-dimensional monitoring data is stored in a local cache according to the time series, and a timestamp is added to each data block.

[0007] As a further embodiment of the present invention, the generation of the anomaly information packet containing node identifiers, anomaly type, and anomaly strength value specifically includes: Real-time calculation of time-domain and frequency-domain characteristic parameters of multi-dimensional monitoring data. The time-domain characteristic parameters include the effective value of current, peak voltage and root mean square value of vibration, and the frequency-domain characteristic parameters include the harmonic content of current, the amplitude of the main frequency of sound wave and the energy of vibration spectrum. The calculated time-domain and frequency-domain characteristic parameters are compared with the preset fault thresholds one by one. When any characteristic parameter continuously exceeds the corresponding fault threshold and the duration exceeds the safety judgment time, it is determined that a line abnormal event of the type corresponding to that characteristic parameter has occurred, and the abnormal intensity value is calculated. The sensor node that detects a line anomaly event is marked as an abnormal node, and the abnormal status flag is set in the node status register. Based on the determination results of line anomaly events, an anomaly information packet containing node identifier, anomaly type, and anomaly intensity value is generated.

[0008] As a further aspect of the present invention, the generation of verification information, including node coordinates, voting type, and confidence weight, specifically includes: Starting with the abnormal node as the initiating node, the abnormal information packet is broadcast to all neighboring nodes within the communication range, and a timer is started to record the broadcast time. The neighboring node that receives the abnormal information packet acts as a verification node and retrieves multi-dimensional monitoring data corresponding to the timestamp of the abnormal information packet from the local cache, including current waveform data, voltage waveform data, vibration acceleration data and sound wave signal data; The verification node extracts the feature parameters corresponding to the anomaly type, calculates the time-domain correlation coefficient between the node and the anomaly intensity value, and analyzes the spectral characteristics to calculate the frequency-domain similarity between the node and the typical fault spectrum corresponding to the anomaly type. The verification conclusion is generated based on the weighted average of the time-domain correlation coefficient and the frequency-domain similarity. When the weighted average is greater than 0.75, the verification information is generated. Otherwise, the verification information is generated. At the same time, the confidence weight in the range of 0-1 is generated according to the magnitude of the weighted average. The confidence weight of the verification information is set to a positive value, and the confidence weight of the verification information is set to a negative value. The complete verification information package is then generated.

[0009] As a further aspect of the present invention, the determination of reaching a fault consensus specifically includes: Set the maximum propagation range, establish a monitoring area centered on the initiating node and covering all reachable nodes within the large propagation range, verify the node forwards the received verification information within the communication range, and decrease the count of the monitoring area range during the forwarding process; The initiating node collects verification information sent by all verification nodes within the monitoring area within a preset time window, including node coordinates, voting type, and confidence weight; The collected verification information is filtered, and valid verification information containing complete node coordinates, voting type, and confidence weight is retained; Calculate the sum of confidence weights for all valid verification information and compare the sum of confidence weights with a consensus threshold that is dynamically adjusted based on historical data. If the sum of confidence weights exceeds the dynamic consensus threshold, it is determined that a fault consensus has been reached; otherwise, it is determined that a fault consensus has not been reached.

[0010] As a further aspect of the present invention, when a fault consensus is reached: Extract the node coordinates and confidence weights of the verification nodes from all verification information, and perform a weighted calculation on the node coordinates and confidence weights of each verification node to obtain the initial fault location coordinates; The fault location coordinates are corrected according to the fault type, and the final fault location coordinates are output.

[0011] As a further aspect of the present invention, when a fault consensus is not reached: When a consensus on the fault cannot be reached, the fault information packets and verification information stored in the faulty nodes and verification nodes are automatically cleared. Update the false alarm statistics record and adaptively adjust the fault threshold parameters in the anomaly detection algorithm based on the false alarm statistics record.

[0012] Another object of the present invention is to provide a power transmission line fault detection and location system based on a sensor network, the system comprising: A multi-dimensional monitoring data acquisition module is used to continuously collect multi-dimensional monitoring data of the transmission line through sensors. The multi-dimensional monitoring data includes electrical quantity data and physical quantity data. The independent abnormal event judgment module is used to independently and in real time determine whether a line abnormality has occurred based on multi-dimensional monitoring data. When any sensor detects a line abnormal event, the sensor node is marked as an abnormal node, and an abnormal information packet containing node identifier, abnormality type and abnormality intensity value is generated. The abnormal information broadcasting module is used to broadcast abnormal information packets to all neighboring nodes within the communication range, with the abnormal node as the initiating node. The neighboring nodes that receive the abnormal information packets act as verification nodes, and perform local verification of the abnormal information packets based on their own collected local multi-dimensional monitoring data. The module calculates the correlation between the multi-dimensional monitoring data and the abnormal information packets, and generates verification information including node coordinates, voting type, and confidence weight. The fault consensus determination module is used to integrate the verification information of all verification nodes within a preset range, with the initiating node as the center. Based on the confidence weight in the verification information, when the sum of the confidence weights exceeds the dynamic consensus threshold, it is determined that a fault consensus has been reached. The fault location calculation and false alarm handling module is used to calculate the fault location coordinates based on the node coordinates and confidence weights of all verification nodes when a fault consensus is reached; when a fault consensus is not reached, it automatically discards all relevant data of the line abnormal event and performs false alarm suppression processing.

[0013] The beneficial effects of this invention are: The present invention provides a sensor network-based method and system for transmission line fault detection and location. It comprehensively captures electrical and physical fault characteristics through multi-dimensional monitoring data acquisition, avoiding missed fault detection due to single data dimensions. Relying on a collaborative mechanism of abnormal node broadcasting and neighbor node verification, combined with dynamic consensus threshold determination, it significantly reduces the risk of false alarms caused by single-node misjudgment and transient interference, thus significantly improving the reliability of fault detection. In the fault location stage, through weighted calculation of confidence weights of verification nodes and targeted correction of fault types, it fully combines data reliability and fault characteristics, effectively reducing location deviation and decreasing the scope and cost of on-site troubleshooting for maintenance personnel. Simultaneously, the invalid data clearing mechanism when fault consensus is not reached releases node storage and computing resources, ensuring the long-term stable operation of the sensor network. Furthermore, the false alarm statistics and adaptive adjustment of fault thresholds enable system self-optimization, enhancing adaptability to different operating environments and conditions, maintaining high-efficiency detection performance over the long term, and ultimately effectively improving the operational stability of transmission lines and the reliability of power supply, providing efficient and accurate technical support for automated monitoring of power systems. Attached Figure Description

[0014] Figure 1 A flowchart of a sensor network-based transmission line fault detection and location method provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating the continuous acquisition of multi-dimensional monitoring data of power transmission lines via sensors, as provided in an embodiment of the present invention. Figure 3 A flowchart for generating an anomaly information packet containing node identifiers, anomaly types, and anomaly strength values ​​is provided for embodiments of the present invention. Figure 4 A flowchart for generating verification information including node coordinates, voting type, and confidence weight, provided for embodiments of the present invention; Figure 5 A flowchart for determining and reaching a consensus on faults is provided in this embodiment of the invention; Figure 6 This is a structural block diagram of a power transmission line fault detection and location system based on a sensor network, provided in an embodiment of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0016] Figure 1 A flowchart of a sensor network-based transmission line fault detection and location method provided in an embodiment of the present invention is shown below. Figure 1 As shown, the method includes: S100 continuously collects multi-dimensional monitoring data of the transmission line through sensors, including electrical quantity data and physical quantity data; Sensors deployed at key locations along transmission lines enable comprehensive, real-time monitoring of the line's operational status. The collected multi-dimensional monitoring data encompasses both electrical and physical quantities. Electrical data focuses on core indicators of power transmission, directly reflecting the stability of the power delivery process. Physical data addresses the line's mechanical structure and external environmental influences, capturing physical changes in the line under different operating conditions.

[0017] During data acquisition, the sensors operate continuously at a preset sampling frequency to ensure no potential precursory signals of faults are missed. The data preprocessing stage addresses potential noise interference and data format differences during acquisition, filtering out invalid interference signals and standardizing data to provide high-quality raw data support for subsequent anomaly detection and verification. The preprocessed data is stored locally in a time series format with timestamps added, establishing temporal correlations between different data dimensions. This ensures accurate tracking of parameter changes within the same time period during subsequent analysis, providing a basis for the time-based localization of anomalies.

[0018] S200, based on multi-dimensional monitoring data, allows each sensor to independently and in real time determine whether a line abnormality has occurred. When any sensor detects a line abnormality event, it marks the sensor node as an abnormal node and generates an abnormal information packet containing a node identifier, abnormality type, and abnormality intensity value. After completing the multi-dimensional monitoring data acquisition and preprocessing, each sensor node independently performs anomaly judgment on the locally stored time-series monitoring data. In real time, it transforms the acquired current and voltage waveforms, vibration acceleration, and acoustic signals into quantifiable time-domain and frequency-domain characteristic parameters. The time-domain characteristic parameters focus on the real-time dynamic characteristics of the data, which can intuitively reflect the current operating status of the line. For example, changes in the effective value of the current can quickly detect whether there is an overload or short-circuit trend in the line, fluctuations in the voltage peak value can promptly detect overvoltage risks, and the root mean square value of vibration can monitor mechanical anomalies such as conductor galloping and broken strands. The frequency-domain characteristic parameters delve into the frequency domain information of the data. Through the current harmonic content, it can identify electrical anomalies caused by nonlinear load access or equipment aging. The amplitude of the main frequency of the acoustic wave can accurately locate the source of faults such as corona discharge and external force damage, and the vibration spectrum energy can detect hidden mechanical faults such as loose hardware and broken insulators.

[0019] These characteristic parameters are compared item by item with the fault thresholds preset based on transmission line design standards, historical fault data and safe operation requirements. The abnormality is determined by the parameter continuously exceeding the corresponding threshold and the duration exceeding the safe judgment time. This filters out instantaneous parameter fluctuations caused by non-fault factors such as lightning strikes and birds briefly touching the line, ensuring the accuracy of abnormality determination.

[0020] When the anomaly detection conditions are met, the sensor node will be marked as an abnormal node, and an abnormal status flag will be set in its own status register so that the node status can be clearly identified during subsequent data interaction and system scheduling. This will generate an anomaly information packet, in which the node identifier can accurately locate the physical location of the node that initiated the anomaly, the anomaly type is directly associated with the fault category to provide direction for subsequent targeted processing, and the anomaly intensity value quantifies the severity of the fault to provide a basis for prioritizing fault processing.

[0021] S300 uses an abnormal node as the initiating node to broadcast an abnormal information packet to all neighboring nodes within the communication range. The neighboring nodes that receive the abnormal information packet act as verification nodes, and perform local verification of the abnormal information packet based on their own collected local multi-dimensional monitoring data. They calculate the correlation between the multi-dimensional monitoring data and the abnormal information packet, and generate verification information including node coordinates, voting type, and confidence weight. An abnormal node, acting as the initiating node, broadcasts an abnormal information packet to all neighboring nodes within its communication range. The broadcast operation synchronously starts a timer to record the broadcast time. The operation of the timer can define the time window for subsequent collection of verification information, ensuring that the verification process is completed within a reasonable timeframe, avoiding delays in fault response due to waiting timeouts, and also filtering out invalid data transmitted after timeouts to ensure the timeliness of verification data.

[0022] The neighboring node that receives the anomaly information packet automatically becomes the verification node. The verification node retrieves multi-dimensional monitoring data from its local cache that corresponds exactly to the timestamp of the anomaly information packet. This data includes current waveform data, voltage waveform data, vibration acceleration data, and acoustic signal data. The accurate correspondence of the timestamps ensures that the retrieved data is completely synchronized with the time period of the anomaly event, eliminating interference from irrelevant data in different time intervals. This allows the verification process to focus on the parameter changes when the anomaly event occurs, improving the targeting of the verification.

[0023] Based on the anomaly type in the anomaly information packet, the verification node extracts the corresponding feature parameters. For example, when the anomaly type is line galloping fault, the verification node will focus on extracting vibration-related feature parameters such as the root mean square value of vibration and vibration spectrum energy. If the anomaly type is corona discharge fault, parameters such as the amplitude of the main frequency of the sound wave will be extracted first.

[0024] Calculate the temporal correlation coefficient between these characteristic parameters and the anomaly intensity value in the anomaly information packet. The temporal correlation coefficient can reflect the consistency of the changing trend of the two in the time dimension. For example, when the effective value of the initiating node current continues to rise, verify whether the synchronous current data of the node shows the same trend. At the same time, analyze the spectral characteristics, calculate the frequency domain similarity with the typical fault spectrum corresponding to the anomaly type, and judge the degree of matching between the two in frequency distribution.

[0025] Verification conclusions are generated based on a weighted average of time-domain correlation coefficients and frequency-domain similarity. The calculation of the weighted average assigns weights according to the differences in the importance of time-domain and frequency-domain features under different fault types. When the weighted average reaches the set standard, verification information is generated; otherwise, verification information is generated. At the same time, confidence weights in the range of 0-1 are generated based on the magnitude of the weighted average. The confidence weights of verification information that passes are set to positive values, and those that fail are set to negative values. Finally, the node coordinates, voting type, and confidence weights are packaged into a complete verification information package.

[0026] S400 integrates the verification information of all verification nodes within a preset range, with the initiating node as the center. Based on the confidence weight in the verification information, when the sum of the confidence weights exceeds the dynamic consensus threshold, it is determined that a fault consensus has been reached. By setting a maximum propagation range and establishing a monitoring area centered on the initiating node that covers all reachable nodes within that range, the verification node forwards the received verification information within the communication range and decrements the monitoring area range during the forwarding process. This design can precisely control the spatial range of nodes participating in the verification, avoid irrelevant nodes from participating due to excessive propagation, and prevent data redundancy and response delays. At the same time, it ensures coverage of all nodes that may be affected by the fault, making the verification more targeted.

[0027] The initiating node will collect the verification information sent by all verification nodes in the monitoring area within a preset time window. The setting of the time window, combined with the communication delay characteristics of the transmission line sensor network, can ensure that all valid verification nodes have enough time to send information, and can also avoid delays in fault response caused by waiting timeout.

[0028] After collection, the verification information is filtered, retaining valid verification information containing complete node coordinates, voting type, and confidence weight, while removing information with missing data or abnormal format to avoid incomplete data interfering with subsequent calculations and ensure the quality of data participating in consensus determination. The sum of the confidence weights of all valid verification information is calculated, and then compared with a consensus threshold dynamically adjusted based on historical data. The adjustment of the dynamic consensus threshold is based on data such as the support of verification nodes for different types of faults and the weight distribution of false alarm events in past fault cases, which can adapt to the differences in fault characteristics under different transmission line environments and operating conditions.

[0029] When a fault consensus is reached, the S500 calculates the fault location coordinates based on the node coordinates and confidence weights of all verification nodes; when a fault consensus is not reached, all relevant data of the line abnormal event is automatically discarded, and false alarm suppression processing is performed.

[0030] The node coordinates and corresponding confidence weights of the verification nodes are extracted from all valid verification information, and the initial fault location coordinates are obtained by weighted summation of the two. Nodes with stronger correlation between data and abnormal events and more credible verification conclusions have a higher weight in the calculation. The confidence weight effectively avoids the weakening of positioning accuracy by nodes that are far from the fault point or whose data has slight interference.

[0031] After obtaining the initial coordinates, targeted corrections are made based on the determined fault type. Different faults have different parameter propagation patterns and impact ranges in transmission lines. Through correction logic that matches the fault type, the final output fault location coordinates are made closer to the actual fault point.

[0032] When a consensus on a fault cannot be reached, the system automatically clears the fault information packets and verification information stored in the abnormal nodes and verification nodes. This prevents invalid data from occupying the nodes' limited storage resources and ensures that each node in the sensor network can continuously and efficiently collect and process monitoring data. Simultaneously, the false alarm statistics record is updated, incorporating the event for which consensus was not reached into the statistical archive. Subsequently, the system will adaptively adjust the fault threshold parameters in the anomaly detection algorithm based on changes in the false alarm statistics record. If the frequency of false alarms caused by a certain characteristic parameter continues to rise, the corresponding fault threshold will be appropriately increased to reduce the probability of subsequent similar instantaneous interference triggering anomaly judgments, allowing the detection standard to dynamically adapt to changes in the line's operating environment.

[0033] like Figure 2 As shown, the continuous acquisition of multi-dimensional monitoring data of the transmission line through sensors includes: The S110 continuously collects current and voltage waveform data, vibration acceleration data, and acoustic signal data of transmission lines through sensors, serving as multi-dimensional monitoring data. S120 preprocesses the collected multi-dimensional monitoring data, stores the preprocessed multi-dimensional monitoring data in a local cache according to the time series, and adds a timestamp to each data block.

[0034] like Figure 3 As shown, generating an anomaly information packet containing node identifiers, anomaly type, and anomaly strength value specifically includes: S210 calculates the time-domain and frequency-domain characteristic parameters of multi-dimensional monitoring data in real time. The time-domain characteristic parameters include the effective value of current, the peak value of voltage, and the root mean square value of vibration. The frequency-domain characteristic parameters include the harmonic content of current, the amplitude of the main frequency of sound wave, and the energy of the vibration spectrum. Time-domain feature parameters: RMS value of current ; Peak voltage ; Root mean square value of vibration ; in: 1: The total number of sampling points for current waveform data is determined by the sensor sampling frequency and data acquisition duration; : No. The instantaneous current value at each sampling point is taken from the current waveform data collected by the sensor; : No. Timestamp of each sampling point; : No. The instantaneous voltage value at each sampling point is taken from the voltage waveform data collected by the sensor; The total number of sampling points for voltage waveform data is consistent with the sampling frequency and acquisition duration for current waveform data. : No. The instantaneous vibration acceleration values ​​at each sampling point are taken from the vibration acceleration data collected by the sensor; The total number of sampling points for vibration acceleration data is determined by the sampling frequency and the acquisition duration.

[0035] Frequency domain characteristic parameters: Current harmonic content ; Sound wave dominant frequency amplitude ; Vibration Spectrum Energy ; in: Harmonic order, ; Maximum harmonic order for analysis; : No. The effective value of the subharmonic current is obtained by performing a fast Fourier transform on the current waveform data; The effective value of the fundamental current is calculated by extracting the fundamental component after FFT decomposition of the current waveform data. Sound wave signals at frequencies The amplitude at that point is obtained by performing an FFT transform on the acoustic signal data; : No. The frequency value of each frequency point is preset to a frequency range of 20Hz-20kHz. The total number of frequency points in the frequency domain data of the acoustic signal is determined by the sampling frequency and the number of FFT points; Vibration acceleration signal at frequency The amplitude at that point is obtained by performing an FFT transform on the vibration acceleration data; The total number of frequency points in the frequency domain data of the vibration signal; Frequency resolution is equal to the ratio of sampling frequency to the number of FFT points.

[0036] S220: The calculated time-domain and frequency-domain characteristic parameters are compared with the preset fault thresholds one by one. When any characteristic parameter continuously exceeds the corresponding fault threshold and the duration exceeds the safety judgment time, it is determined that a line abnormal event of the type corresponding to the characteristic parameter has occurred, and the abnormal intensity value is calculated. ; Among them, when hour, (Only the contribution of feature parameters that exceed the fault threshold is considered).

[0037] in: Anomaly intensity value reflects the severity of abnormal events on the line, and its value range is [range missing]. ; The total number of characteristic parameters involved in the calculation of abnormal intensity, namely the aforementioned 6 characteristic parameters: effective value of current, peak voltage, root mean square value of vibration, harmonic content of current, amplitude of dominant frequency of sound wave, and energy of vibration spectrum; : No. The weight coefficients of each feature parameter are preset based on the importance of the feature parameter in relation to the fault, satisfying the following: ; : No. The measured values ​​of each characteristic parameter, i.e., the time-domain or frequency-domain characteristic parameter values ​​calculated in real time; : No. The fault threshold corresponding to each characteristic parameter is a preset critical value based on the transmission line design standards, historical fault data, and safe operation requirements.

[0038] When any characteristic parameter continuously exceeds the corresponding fault threshold, and the duration exceeds the safety judgment time, a line abnormality event of the corresponding type is determined to have occurred. A real fault in a transmission line will cause the characteristic parameters of the relevant physical or electrical quantities to deviate from the normal range, and this deviation is continuous. After a short-circuit fault occurs, the effective value of the current will rise instantly and remain in the state above the threshold until the fault is cleared. Abnormal line vibration caused by icing will persist as the thickness of the icing increases, rather than fluctuating instantaneously.

[0039] Numerous transient interference factors exist in the operating environment of power transmission lines. These interferences can cause characteristic parameters to briefly exceed thresholds, but they are not actual faults. By setting conditions that persist for a duration exceeding the safety judgment period, such transient interferences can be effectively filtered out, avoiding false alarms.

[0040] S230, mark the sensor node that detects the abnormal event of the line as an abnormal node, and set the abnormal status flag in the node status register; S240, based on the determination result of the line abnormal event, generates an abnormal information packet containing node identifier, abnormal type and abnormal intensity value.

[0041] The determination of the anomaly type is based on the mapping relationship between characteristic parameters exceeding the fault threshold and typical fault types of transmission lines: If multiple feature parameters exceed the fault threshold simultaneously, the anomaly type corresponding to the feature parameter with the highest weight coefficient that exceeds the threshold is taken as the final anomaly type. The name and code of the anomaly type are preset in the sensor. When generating an anomaly information packet, the anomaly type code and name corresponding to the feature parameter that exceeds the standard are directly extracted to ensure consistency and standardization.

[0042] like Figure 4 As shown, the generation of verification information, including node coordinates, voting type, and confidence weight, specifically includes: S310, with the abnormal node as the initiating node, broadcasts the abnormal information packet to all neighboring nodes within the communication range, and starts a timer to record the broadcast time; S320, the neighboring node that receives the abnormal information packet acts as the verification node, and retrieves multi-dimensional monitoring data corresponding to the timestamp of the abnormal information packet from the local cache, including current waveform data, voltage waveform data, vibration acceleration data and sound wave signal data; S330, the verification node extracts the feature parameters corresponding to the anomaly type, calculates the time domain correlation coefficient between the node and the anomaly intensity value, and analyzes the spectral characteristics to calculate the frequency domain similarity between the node and the typical fault spectrum corresponding to the anomaly type. Time-domain correlation coefficient: ; in: The temporal correlation coefficient reflects the temporal correlation between the characteristic parameters of the verification node and the anomaly strength value of the initiating node, and its value range is... The closer the value is to 1, the stronger the correlation. The total number of sampling points in the time series corresponds to the duration of the abnormal event of the initiating node, i.e., the number of sampling points within the security judgment period; The verification node is at the 1st The measured values ​​of the target feature parameters of each sampling point, where the target feature parameters are the feature parameters of the initiating node that exceed the threshold. : Verify the average value of the target feature parameter sequence of the node; The initiating node is at the The anomaly intensity contribution value of each sampling point, i.e. ,in For the initiating node Measured values ​​of target feature parameters at each sampling point; : The average value of the sequence of anomaly strength contribution values ​​of the initiating node.

[0043] Frequency domain similarity: ; in: Frequency domain similarity, dimensionless, reflects the degree of similarity between the characteristic parameter spectrum of the verification node and the spectrum of a typical fault, with a value range of [value missing]. The closer the similarity is to 1, the higher the similarity. The total number of frequency points in the frequency domain data is consistent with the frequency points of the frequency domain analysis of the feature parameters of the initiating node; Typical fault spectrum corresponding to the initiating node anomaly type in frequency The amplitude at the location, and the typical fault spectrum is a standard spectrum template pre-stored based on historical fault data.

[0044] S340: Generate verification conclusions based on the weighted average of time-domain correlation coefficient and frequency-domain similarity. When the weighted average is greater than 0.75, generate verification pass information; otherwise, generate verification fail information. At the same time, generate confidence weights in the range of 0-1 based on the magnitude of the weighted average. The confidence weight of verification pass information is set to a positive value, and the confidence weight of verification fail information is set to a negative value. Package the results into a complete verification information package.

[0045] The verification information indicates that the verification node, based on its own data, confirms / supports the abnormal event reported by the initiating node, confirming the strong correlation between its own data and the abnormal event reported by the initiating node. This indicates that the changes in the characteristic parameters of the area where the verification node is located are consistent with the abnormal characteristics of the initiating node, and its data characteristics are highly consistent with the abnormal event, making a positive contribution to the formation of fault consensus. Therefore, the confidence weight is set to a positive value. A failed verification message indicates that the verification node, based on its own data, could not confirm or refute the anomaly. This means the verification node determined that its own data had no correlation or a very weak correlation with the anomaly reported by the initiating node. It suggests that no parameter changes matching the anomaly characteristics of the initiating node occurred in the verification node's region, and the anomaly may be a false alarm or local interference from the initiating node. The data characteristics do not match the anomaly, indicating that the authenticity of the anomaly is questionable. Therefore, the confidence weight is set to a negative value.

[0046] like Figure 5 As shown, the determination of reaching a fault consensus specifically includes: S410 sets the maximum propagation range, establishes a monitoring area centered on the initiating node and covering all reachable nodes within the large propagation range, verifies that the node forwards the received verification information within the communication range, and decrements the monitoring area range during the forwarding process; S420: The initiating node collects verification information sent by all verification nodes within the monitoring area within a preset time window, including node coordinates, voting type, and confidence weight. S430 filters the collected verification information and retains valid verification information containing complete node coordinates, voting type, and confidence weight; S440: Calculate the sum of confidence weights for all valid verification information and compare the sum of confidence weights with a consensus threshold that is dynamically adjusted based on historical data. If the sum of confidence weights exceeds the dynamic consensus threshold, it is determined that a fault consensus has been reached; otherwise, it is determined that a fault consensus has not been reached.

[0047] A power transmission line fault can affect multiple sensor nodes within a certain range. The verification information from multiple neighboring nodes can cross-verify the authenticity of the abnormal event from a spatial dimension. The sum of confidence weights reflects the overall support of multiple nodes for the abnormal event. The higher the sum, the stronger the authenticity of the fault.

[0048] The total confidence weights follow a normal distribution. In the event of a real fault, the proportion of supporting verification nodes (nodes that send verification information that has passed) is high, and the total weights are likely to exceed the threshold. In the event of a false alarm, the proportion of opposing verification nodes (nodes that send verification information that has failed) is high, and the total weights are likely to be lower than the threshold.

[0049] When a consensus on the failure is reached: Extract the node coordinates and confidence weights of the verification nodes from all verification information, and perform a weighted calculation on the node coordinates and confidence weights of each verification node to obtain the initial fault location coordinates; The fault location coordinates are corrected according to the fault type, and the final fault location coordinates are output.

[0050] When a consensus on the failure is not reached: When a consensus on the fault cannot be reached, the fault information packets and verification information stored in the faulty nodes and verification nodes are automatically cleared. Update the false alarm statistics record and adaptively adjust the fault threshold parameters in the anomaly detection algorithm based on the false alarm statistics record.

[0051] Figure 6 The structural block diagram of the power transmission line fault detection and location system based on sensor network provided in the embodiments of the present invention is as follows: Figure 6 As shown, the system includes: The multi-dimensional monitoring data acquisition module 100 is used to continuously acquire multi-dimensional monitoring data of the transmission line through sensors. The multi-dimensional monitoring data includes electrical quantity data and physical quantity data. The independent judgment module 200 for abnormal events is used to independently and in real time determine whether a line abnormality has occurred based on multi-dimensional monitoring data. When any sensor detects a line abnormality event, the sensor node is marked as an abnormal node, and an abnormal information packet containing node identifier, abnormality type and abnormality intensity value is generated. The abnormal information broadcasting module 300 is used to broadcast abnormal information packets to all neighboring nodes within the communication range, with the abnormal node as the initiating node. The neighboring nodes that receive the abnormal information packets act as verification nodes, perform local verification of the abnormal information packets based on their own collected local multi-dimensional monitoring data, calculate the correlation between the multi-dimensional monitoring data and the abnormal information packets, and generate verification information including node coordinates, voting type and confidence weight. The fault consensus determination module 400 is used to integrate the verification information of all verification nodes within a preset range, with the initiating node as the center. Based on the confidence weight in the verification information, when the sum of the confidence weights exceeds the dynamic consensus threshold, it is determined that a fault consensus has been reached. The fault location calculation and false alarm handling module 500 is used to calculate the fault location coordinates based on the node coordinates and confidence weights of all verification nodes when a fault consensus is reached; when a fault consensus is not reached, it automatically discards all relevant data of the line abnormal event and performs false alarm suppression processing.

[0052] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0053] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for fault detection and location of transmission lines based on sensor networks, characterized in that, The method includes: The transmission line continuously collects multi-dimensional monitoring data through sensors, including electrical quantity data and physical quantity data. Based on multi-dimensional monitoring data, each sensor independently and in real time determines whether a line abnormality has occurred. When any sensor detects a line abnormality event, it marks the sensor node as an abnormal node and generates an abnormal information packet containing a node identifier, abnormality type, and abnormality intensity value. An abnormal node is used as the initiating node to broadcast an abnormal information packet to all neighboring nodes within the communication range. The neighboring nodes that receive the abnormal information packet act as verification nodes, and perform local verification of the abnormal information packet based on their own collected local multi-dimensional monitoring data. They calculate the correlation between the multi-dimensional monitoring data and the abnormal information packet, and generate verification information including node coordinates, voting type and confidence weight. Centered on the initiating node, the verification information of all verification nodes within a preset range is integrated. Based on the confidence weight in the verification information, when the sum of the confidence weights exceeds the dynamic consensus threshold, a fault consensus is determined to be reached. When a fault consensus is reached, the fault location coordinates are calculated based on the node coordinates and confidence weights of all verification nodes; when a fault consensus is not reached, all relevant data of the line abnormal event are automatically discarded, and false alarm suppression processing is performed.

2. The method according to claim 1, characterized in that, The continuous acquisition of multi-dimensional monitoring data of transmission lines via sensors includes: The system continuously collects current and voltage waveform data, vibration acceleration data, and acoustic signal data of the transmission line through sensors, serving as multi-dimensional monitoring data. The collected multi-dimensional monitoring data is preprocessed, and the preprocessed multi-dimensional monitoring data is stored in a local cache according to the time series, and a timestamp is added to each data block.

3. The method according to claim 2, characterized in that, The generation of the anomaly information packet, which includes node identifiers, anomaly type, and anomaly strength value, specifically includes: Real-time calculation of time-domain and frequency-domain characteristic parameters of multi-dimensional monitoring data. The time-domain characteristic parameters include the effective value of current, peak voltage and root mean square value of vibration, and the frequency-domain characteristic parameters include the harmonic content of current, the amplitude of the main frequency of sound wave and the energy of vibration spectrum. The calculated time-domain and frequency-domain characteristic parameters are compared with the preset fault thresholds one by one. When any characteristic parameter continuously exceeds the corresponding fault threshold and the duration exceeds the safety judgment time, it is determined that a line abnormal event of the type corresponding to that characteristic parameter has occurred, and the abnormal intensity value is calculated. The sensor node that detects a line anomaly event is marked as an abnormal node, and the abnormal status flag is set in the node status register. Based on the determination results of line anomaly events, an anomaly information packet containing node identifier, anomaly type, and anomaly intensity value is generated.

4. The method according to claim 3, characterized in that, The generated verification information includes node coordinates, voting type, and confidence weight, specifically including: Starting with the abnormal node as the initiating node, the abnormal information packet is broadcast to all neighboring nodes within the communication range, and a timer is started to record the broadcast time. The neighboring node that receives the abnormal information packet acts as a verification node and retrieves multi-dimensional monitoring data corresponding to the timestamp of the abnormal information packet from the local cache, including current waveform data, voltage waveform data, vibration acceleration data and sound wave signal data; The verification node extracts the feature parameters corresponding to the anomaly type, calculates the time-domain correlation coefficient between the node and the anomaly intensity value, and analyzes the spectral characteristics to calculate the frequency-domain similarity between the node and the typical fault spectrum corresponding to the anomaly type. The verification conclusion is generated based on the weighted average of the time-domain correlation coefficient and the frequency-domain similarity. When the weighted average is greater than 0.75, the verification information is generated. Otherwise, the verification information is generated. At the same time, the confidence weight in the range of 0-1 is generated according to the magnitude of the weighted average. The confidence weight of the verification information is set to a positive value, and the confidence weight of the verification information is set to a negative value. The complete verification information package is then generated.

5. The method according to claim 4, characterized in that, The determination of reaching a fault consensus specifically includes: Set the maximum propagation range, establish a monitoring area centered on the initiating node and covering all reachable nodes within the large propagation range, verify the node forwards the received verification information within the communication range, and decrease the count of the monitoring area range during the forwarding process; The initiating node collects verification information sent by all verification nodes within the monitoring area within a preset time window, including node coordinates, voting type, and confidence weight; The collected verification information is filtered, and valid verification information containing complete node coordinates, voting type, and confidence weight is retained; Calculate the sum of confidence weights for all valid verification information and compare the sum of confidence weights with a consensus threshold that is dynamically adjusted based on historical data. If the sum of confidence weights exceeds the dynamic consensus threshold, it is determined that a fault consensus has been reached; otherwise, it is determined that a fault consensus has not been reached.

6. The method according to claim 5, characterized in that, When a consensus on the failure is reached: Extract the node coordinates and confidence weights of the verification nodes from all verification information, and perform a weighted calculation on the node coordinates and confidence weights of each verification node to obtain the initial fault location coordinates; The fault location coordinates are corrected according to the fault type, and the final fault location coordinates are output.

7. The method according to claim 5, characterized in that, When a consensus on the failure is not reached: When a consensus on the fault cannot be reached, the fault information packets and verification information stored in the faulty nodes and verification nodes are automatically cleared. Update the false alarm statistics record and adaptively adjust the fault threshold parameters in the anomaly detection algorithm based on the false alarm statistics record.

8. A power transmission line fault detection and location system based on sensor networks, characterized in that, The system includes: A multi-dimensional monitoring data acquisition module is used to continuously collect multi-dimensional monitoring data of the transmission line through sensors. The multi-dimensional monitoring data includes electrical quantity data and physical quantity data. The independent abnormal event judgment module is used to independently and in real time determine whether a line abnormality has occurred based on multi-dimensional monitoring data. When any sensor detects a line abnormal event, the sensor node is marked as an abnormal node, and an abnormal information packet containing node identifier, abnormality type and abnormality intensity value is generated. The abnormal information broadcasting module is used to broadcast abnormal information packets to all neighboring nodes within the communication range, with the abnormal node as the initiating node. The neighboring nodes that receive the abnormal information packets act as verification nodes, and perform local verification of the abnormal information packets based on their own collected local multi-dimensional monitoring data. The module calculates the correlation between the multi-dimensional monitoring data and the abnormal information packets, and generates verification information including node coordinates, voting type, and confidence weight. The fault consensus determination module is used to integrate the verification information of all verification nodes within a preset range, with the initiating node as the center. Based on the confidence weight in the verification information, when the sum of the confidence weights exceeds the dynamic consensus threshold, it is determined that a fault consensus has been reached. The fault location calculation and false alarm handling module is used to calculate the fault location coordinates based on the node coordinates and confidence weights of all verification nodes when a fault consensus is reached; when a fault consensus is not reached, it automatically discards all relevant data of the line abnormal event and performs false alarm suppression processing.

Images