Rapid fault positioning system for power distribution network

By combining Kalman filtering, Bayesian network and convolutional neural network in the rapid positioning system of distribution network faults, rapid positioning and accurate identification of distribution network faults is achieved, and the long fault processing time and data in the existing technology are solved The lack of analysis has improved the response speed and management efficiency of the power system.

CN118980884AActive Publication Date: 2024-11-19GONGYI POWER SUPPLY CO OF STATE GRID HENAN ELECTRIC POWER CO +1

Patent Information

Application Number
CN202410977423.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-19
Estimated Expiration
2044-07-22

AI Technical Summary

Technical Problem

The existing distribution network fault detection and positioning methods rely on traditional monitoring equipment and manual diagnosis, resulting in extended fault processing time, difficult data synchronization, insufficient analysis of fault data, affecting efficiency and accuracy.

Method used

A fast positioning system for faults in the distribution network was designed, including environmental perception module, data fusion processing module, fault detection positioning module, synchronization association module and spatial information integration module, using improved Kalman filtering, Bayesian network and wavelet transformation and convolutional neural networks, combined with geographic information systems, realize time-frequency analysis and intuitive visualization of fault signals.

Benefits of technology

It improves the speed and accuracy of fault location, shortens the power outage time, optimizes the fault handling process, and enhances the reliability and management capabilities of the power system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118980884A_ABST
    Figure CN118980884A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of power grid fault positioning, in particular to a power distribution network fault rapid positioning system which comprises an environment sensing module, a data fusion processing module, a fault detection positioning module, a synchronous association module and a spatial information integration module. Wherein the environment sensing module is used for monitoring the operating environment of the power distribution network in real time; the data fusion processing module is used for extracting effective information related to the fault; the fault detection and positioning module is used for extracting fault characteristic signals and carrying out identification and preliminary positioning; the synchronous association module is used for optimizing an output result of the fault detection positioning module; and the spatial information integration module is used for generating a visual fault distribution diagram. According to the invention, through integration of advanced signal processing technology, time synchronization, event correlation analysis and geographic information system visualization, the fault detection speed and accuracy are significantly improved, the fault information processing flow is optimized, and the operation and maintenance efficiency and the power grid management efficiency are improved through visual fault display.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of power grid fault location, and in particular to a distribution network fault rapid location system. Background Art

[0002] In modern power systems, the stable operation of distribution networks is of vital importance. Rapid detection and accurate positioning of faults are the key to ensuring power supply and system safety. At present, the identification and positioning of distribution network faults mainly rely on traditional monitoring equipment and manual diagnosis. These methods often rely on a large amount of sensor data and complex manual analysis, which leads to prolonged fault handling time and difficulty in meeting the needs of rapid response. In addition, common problems in existing technologies include difficulty in data synchronization, insufficient fault data analysis, and non-intuitive display of fault information, which seriously affect the efficiency and accuracy of fault handling.

[0003] In order to solve the above problems, it is necessary to develop a new fault detection and location system that can synchronize the time data of each node in the distribution network, analyze the fault signal using advanced data processing technology, and provide intuitive fault distribution visualization by integrating geographic information system data, thereby improving the speed and accuracy of fault location. Summary of the invention

[0004] Based on the above purpose, the present invention provides a distribution network fault rapid location system.

[0005] The distribution network fault rapid location system includes an environment perception module, a data fusion processing module, a fault detection and location module, a synchronization association module and a spatial information integration module; wherein:

[0006] Environmental perception module: used to monitor the distribution network operating environment in real time, including temperature, humidity, meteorological conditions and electromagnetic interference parameters;

[0007] Data fusion processing module: Receives the operating environment data transmitted by the environmental perception module, as well as the electrical quantity data of each node in the distribution network, processes the multi-source data through an algorithm combining improved Kalman filtering and Bayesian network, and extracts effective information related to the fault;

[0008] Fault detection and positioning module: Based on the effective information provided by the data fusion processing module, the improved wavelet transform is first used to perform time-frequency analysis on the signal to extract the fault characteristic signal; then the convolutional neural network is combined to identify the fault type and preliminarily locate it, and finally the detection results and positioning information are transmitted to the synchronous association module;

[0009] Synchronous correlation module: receives the output results of the fault detection and positioning module, synchronizes the time data of each node in the distribution network, and uses event correlation technology to analyze the temporal and spatial relationship of fault events, thereby optimizing the output results of the fault detection and positioning module;

[0010] Spatial information integration module: receives the fault location information optimized by the synchronous association module, integrates it with the geographic information system data, and generates a visual fault distribution map.

[0011] Furthermore, the environment perception module includes a temperature monitoring unit, a humidity monitoring unit, a meteorological condition monitoring unit and an electromagnetic interference monitoring unit; wherein:

[0012] Temperature monitoring unit: includes multiple temperature sensors, which are used to collect temperature data of different nodes in the distribution network in real time. The temperature sensor is based on the thermocouple principle and uses the electromotive force difference generated by metal at different temperatures to measure the temperature. The digital signal processor converts the analog signal into a digital signal and transmits it to the data fusion processing module in real time.

[0013] Humidity monitoring unit: includes multiple humidity sensors, which are used to collect humidity data of different nodes of the distribution network in real time. The humidity sensor is based on the principle of capacitance humidity measurement, and uses the change of capacitance value caused by humidity change to measure humidity. The signal is amplified and filtered through the signal conditioning circuit, and then transmitted to the data fusion processing module;

[0014] Meteorological condition monitoring unit: including multiple meteorological sensors for real-time collection of meteorological data in the distribution network area, including wind speed, wind direction, rainfall and atmospheric pressure; the meteorological sensors are based on mechanical and electronic measurement principles, measure meteorological parameters through mechanical movement and electronic signal conversion, and transmit the collected data to the data fusion processing module;

[0015] Electromagnetic interference monitoring unit: includes multiple electromagnetic interference sensors, which are used to detect electromagnetic interference signals in the distribution network in real time. The electromagnetic interference sensor is based on the principle of electromagnetic induction and uses the induced electromotive force generated by the induction coil when the electromagnetic field changes to detect the electromagnetic interference signal. After being processed by filtering and amplifying circuits, the data is transmitted to the data fusion processing module.

[0016] Furthermore, the data fusion processing module includes a data receiving unit, a preprocessing unit, an improved Kalman filter unit, a Bayesian network processing unit and a data output unit; wherein:

[0017] Data receiving unit: used to receive the operating environment data from the environment perception module and the electrical quantity data of each node in the distribution network. The data receiving unit is connected to the environment perception module and the sensor interface of each node in the distribution network through a high-speed data bus to ensure real-time and accurate data transmission;

[0018] Preprocessing unit: performs preliminary processing on the received multi-source data, including data format conversion, noise filtering and data alignment;

[0019] Improved Kalman filter unit: Dynamically filter and estimate the state of preprocessed data. The improved Kalman filter unit is based on the extended Kalman filter algorithm and combines the dynamic model of the distribution network to perform state prediction and error correction on the operating environment data and electrical quantity data to extract preliminary fault-related information;

[0020] Bayesian network processing unit: performs probability inference and data fusion on the preliminary information output by the Kalman filter unit. Specifically, it uses the conditional probability distribution model of the Bayesian network to fuse multi-source data, and combines historical fault data and prior knowledge to deeply extract effective information related to the fault.

[0021] Data output unit: Send the fused data to the downstream modules in real time through the preset standard communication protocol and interface.

[0022] Furthermore, the improved Kalman filter unit includes:

[0023] State initialization: define the initial state variable x of the system 0 and the initial covariance matrix P 0 , used to describe the initial state and initial uncertainty of the system;

[0024] Prediction step: According to the dynamic model of the distribution network, the system state is predicted through the state transition equation, and the process noise covariance Q is used 1 Correct the prediction results; the specific calculation formula is:

[0025] x k|k-1 =f(x k-1 ,u k-1 );

[0026] Among them, x k|k-1 is the predicted state from time k-1 to time k; u k-1 is the control input at time k-1; F 1 is the Jacobian matrix of the state transfer matrix; Q 1 is the process noise covariance;

[0027] Measurement update step: Correct the predicted state according to the measurement equation, and correct the state by measuring the predicted value and the Kalman gain. The specific correction formula is:

[0028] z k|k-1 =h(x k|k-1 );

[0029]

[0030] Among them, zk|k-1 is the measurement prediction value vector, which represents the measurement value calculated according to the predicted state; H 1 is the Jacobian matrix of the measurement matrix, which represents the linear approximation of the measurement model; S 1 is the measurement prediction error covariance matrix, which represents the covariance of the measurement error; R 1 is the measurement noise covariance matrix, representing the measurement noise; K 1 is the Kalman gain matrix, which represents the weight of the prediction error correction;

[0031] State update: Use the Kalman gain to correct the predicted state to obtain the updated state estimate and covariance matrix. The specific formula is:

[0032] x k =x k|k-1 +K 1 (z k -z k|k-1 );

[0033] Among them, x k is the updated state estimation vector, which represents the corrected system state; P k is the updated covariance matrix, which represents the corrected state uncertainty; z k is the actual measurement value vector, which represents the actual data obtained from the environment perception module and the distribution network nodes;

[0034] Fault information extraction: Estimate x based on the updated state k and the covariance matrix P k , extract preliminary fault-related information.

[0035] Furthermore, the Bayesian network processing unit includes:

[0036] Node definition: Define the nodes in the Bayesian network, where each node represents a random variable, specifically the state estimate x output by the Kalman filter unit k 、Historical fault data k and prior knowledge p k ;

[0037] Edge definition: Define the edges in the Bayesian network to represent the dependency between nodes, and use the conditional probability distribution model to describe the dependency;

[0038] Conditional probability table construction: Based on historical data and expert knowledge, a conditional probability table is constructed to describe the probability distribution of each node under a given parent node. k Given the historical fault data h k and prior knowledge p kThe conditional probability in this case can be expressed as: P(x k =x|h k =h,p k =p);

[0039] Probabilistic inference: preliminary information x on the output of the Kalman filter unit k Probabilistic inference is performed, and the posterior probability of a fault is calculated by combining historical fault data and prior knowledge. The Bayesian inference formula is:

[0040] Among them, z k is the actual measurement value obtained from the environment perception module and the distribution network node; P(z k |x k ) is the likelihood function; P(x k ) is the prior probability; P(z k ) as evidence;

[0041] Data fusion: The preliminary information is fused using the Bayesian network, and the conditional probability distribution model and historical data are combined to deeply extract the fault-related information. The fused information includes the updated fault probability distribution and fault location estimation. The fusion processing formula is: Among them, F k is the fused fault information; is the i-th state estimate; is the estimated posterior probability of the i-th state.

[0042] Furthermore, the fault detection and positioning module includes a signal preprocessing unit, a time-frequency analysis unit, a fault identification unit, a positioning unit and a result transmission unit; wherein:

[0043] Signal preprocessing unit: used to receive the effective information provided by the data fusion processing module and preprocess the signal, including denoising and normalization, to improve signal quality and analysis accuracy;

[0044] Time-frequency analysis unit: Use improved wavelet transform to perform time-frequency analysis on the pre-processed signal. Through the preset mother wavelet function of the distribution network signal characteristics, the time domain signal is converted to the time-frequency domain, making the fault characteristics more obvious at different frequencies and time scales, thereby improving the resolution and recognition ability of the fault signal to capture the transient characteristics of the fault signal;

[0045] Fault identification unit: Combines convolutional neural network to identify the fault type of the extracted fault feature signal. Specifically, the fault feature signal is input into the trained convolutional neural network model, and through multi-layer convolution and pooling operations, high-level features are extracted and classified to identify the fault type.

[0046] Positioning unit: Based on the fault type information output by the fault identification unit, the fault is initially located. By analyzing the time-frequency characteristics and spatial distribution information of the fault signal, the specific location of the fault is determined. The positioning information and fault type information are integrated to generate the fault detection result.

[0047] Result transmission unit: transmits fault detection results and location information to the synchronization association module.

[0048] Furthermore, the time-frequency analysis unit includes:

[0049] Select mother wavelet function: According to the characteristics of distribution network fault signal, select a suitable mother wavelet function ψ(t), which includes Daubechies wavelet, Haar wavelet and Symlet wavelet;

[0050] Signal decomposition: The preprocessed time domain signal x(t) is decomposed into different frequencies and time scales through wavelet transform. The formula of continuous wavelet transform is: Among them, W x (a, b) are wavelet coefficients, representing the signal component at scale a and position b; ψ is the mother wavelet function; a is the scale parameter; b is the translation parameter; ψ * is the complex conjugate of the mother wavelet function;

[0051] Multi-scale analysis: By adjusting the scale parameter a and the translation parameter b, the signal is analyzed at multiple scales. The small scale corresponds to high-frequency details, which is used to capture the transient characteristics of the signal; the large scale corresponds to low-frequency information, which is used to reflect the overall trend of the signal. The distribution of fault characteristics at different scales and times can be observed through the wavelet coefficient graph;

[0052] Feature extraction: By analyzing the wavelet coefficient graph, the fault feature signal is extracted. For transient fault signals, the feature will appear as a sudden change in the high-frequency component. The energy or amplitude of the wavelet coefficient is calculated to quantify the fault feature.

[0053] Characteristic signal reconstruction: According to the extracted characteristic signal, the wavelet coefficients are inversely transformed to reconstruct the fault characteristic signal. The reconstruction formula of discrete wavelet transform is: x(t) = ∑ a,b W x (a,b)ψ a,b (t), where ψ a,b (t) is the mother wavelet function at scale a and position b.

[0054] Furthermore, the positioning unit includes:

[0055] Receiving fault type information: The positioning unit first receives the fault type information from the fault identification unit, including the fault type and its corresponding signal feature description;

[0056] Characteristic signal reconstruction: Reconstruct the fault characteristic signal to obtain specific time-frequency characteristics and spatial distribution information;

[0057] Time-frequency analysis of fault signals: Analyze the time-frequency characteristics of the reconstructed signal, especially the transient characteristics, including spikes and breaks. This characteristic is more obvious in the high-frequency components of the wavelet transform. By quantitatively analyzing the intensity and distribution of the transient characteristics, the time point of the fault occurrence can be determined;

[0058] Spatial distribution analysis: Combine the topology of the distribution network and the known sensor locations to analyze the propagation and attenuation of fault characteristic signals at different locations. Use the speed and direction of signal propagation and the signal strength received from different sensors to calculate the potential location of the fault source.

[0059] Accurately locate the fault location: Based on the results of time-frequency feature analysis and spatial distribution analysis, the least squares estimation or triangulation positioning algorithm is used to accurately calculate the specific location of the fault;

[0060] Integrate fault information and output: Integrate location information and fault type information to generate fault detection results, including fault type, specific time and location of the fault, and cause of the fault.

[0061] Furthermore, the synchronization association module includes a time synchronization unit, an event extraction unit, a spatiotemporal association analysis unit, and a result optimization unit; wherein:

[0062] Time synchronization unit: used to receive the output results from the fault detection and positioning module, as well as the time data of each node in the distribution network, and achieve time synchronization through the preset network time protocol to ensure that the timestamps of all nodes are consistent, eliminate clock drift and time error, and provide an accurate time reference;

[0063] Event extraction unit: extracts relevant event information from the output results of the fault detection and positioning module, including the time, type and preliminary positioning information of the fault. The event extraction unit analyzes the fault signal and identifies transient events related to the fault, providing basic data for subsequent spatiotemporal correlation analysis.

[0064] Time-space correlation analysis unit: Based on the synchronization time data provided by the time synchronization unit and the event information extracted by the event extraction unit, the event correlation technology is used to analyze the time-space relationship of the fault event. The specific analysis includes time series analysis.

[0065] The timing analysis uses the synchronous time data to perform timing analysis on the fault events, determine the sequence and time interval of the events, and identify the fault propagation path and speed by comparing the timestamps of different nodes. The timing analysis formula is: Where Δt is the time difference between node i and node j, and is the timestamp of the corresponding node;

[0066] The spatial correlation analysis combines the geographic information system data of the distribution network to analyze the spatial distribution and correlation of fault events, and uses the geographical location of the event and the sensor layout to determine the source of the fault and the scope of influence. The spatial correlation analysis formula is: Where d is the distance between two nodes, (x 1 ,y 1 ) and (x 2 ,y 2 ) is the position coordinate of the node;

[0067] Result optimization unit: According to the analysis results of the time-space correlation analysis unit, the preliminary positioning information of the fault detection and positioning module is optimized. Specifically, the preliminary positioning results are corrected by comprehensively analyzing the time sequence and spatial distribution of the fault events, and the positioning accuracy is improved. The optimized positioning information and fault type information are integrated to generate the final fault detection result. The optimization positioning formula is: Among them, L optimized is the optimized positioning result, L initial is the initial positioning result, L i is the location information of the i-th node, α and β are weighting coefficients.

[0068] Furthermore, the spatial information integration module includes a data receiving unit, a data conversion unit, a spatial data processing unit and a visualization generation unit; wherein:

[0069] Data receiving unit: used to receive optimized fault location information from the synchronization association module and geographic data obtained from the geographic information system;

[0070] Data conversion unit: converts the received optimized fault location information into a format that can be recognized by the geographic information system, and uses the spatial reference system and projection conversion method to match the location information with the geographic coordinates to ensure accurate mapping of the fault location information in the geographic space;

[0071] Spatial data processing unit: integrates the converted fault location information with the geographic information system data. The spatial data processing unit uses spatial analysis technology to process and analyze the fault information, including spatial overlay, buffer analysis and spatial interpolation operations, to generate accurate fault distribution information;

[0072] Visualization generation unit: Based on the integrated processed fault location information, a visual fault distribution map is generated. The visualization generation unit uses the graphics rendering function of the GIS software to display the fault information in a graphical manner, including information on the fault point, impact range and fault type.

[0073] Beneficial effects of the present invention:

[0074] The present invention, by integrating improved wavelet transform and convolutional neural network (CNN), can effectively extract key fault features from a large amount of complex sensor data. This method uses the multi-scale analysis capability of wavelet transform to capture transient fault signals, while CNN conducts in-depth learning and classification of these features, greatly improving the accuracy and speed of fault detection. Through this efficient signal processing and fault analysis technology, the system can respond to faults quickly, shorten power outage time, and improve the reliability of the power system.

[0075] The present invention, by introducing a time synchronization unit and spatiotemporal correlation analysis technology, ensures that the data of each node in the distribution network are highly synchronized in time, thereby making event correlation analysis more accurate. This synchronization and analysis system can clearly identify the propagation path and speed of the fault, as well as the specific location where the fault occurs, which not only helps to quickly locate the source of the fault, but also provides important information for subsequent maintenance work, thereby optimizing the fault handling process and improving repair efficiency.

[0076] The present invention, by utilizing geographic information system technology and combining fault location information, generates an intuitive fault distribution map. This visualization not only enables operation and maintenance personnel to directly observe the geographical location and impact range of the fault, but also enables comparative analysis with historical data and geographical environment to further assist in decision making. By providing this intuitive fault display method, the system greatly improves the efficiency of fault response and processing, and also strengthens the management and monitoring capabilities of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0078] Figure 1 A schematic diagram of a system for quickly locating faults in a power distribution network according to an embodiment of the present invention;

[0079] Figure 2 Schematic diagram of a fault detection and positioning module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0080] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0081] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. The words "include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0082] like Figure 1-Figure 2 As shown, the distribution network fault rapid location system includes an environment perception module, a data fusion processing module, a fault detection and location module, a synchronization association module and a spatial information integration module; wherein:

[0083] Environmental perception module: used to monitor the distribution network operating environment in real time, including temperature, humidity, meteorological conditions and electromagnetic interference parameters, and transmit these data to the data fusion processing module;

[0084] Data fusion processing module: Receives the operating environment data transmitted by the environmental perception module, as well as the electrical quantity data of each node in the distribution network, processes the multi-source data through an algorithm combining improved Kalman filtering and Bayesian network, extracts effective information related to the fault, and transmits it to the fault detection and location module;

[0085] Fault detection and positioning module: Based on the effective information provided by the data fusion processing module, the improved wavelet transform is first used to perform time-frequency analysis on the signal to extract the fault characteristic signal; then the convolutional neural network is combined to identify the fault type and preliminarily locate it, and finally the detection results and positioning information are transmitted to the synchronous association module;

[0086] Synchronous correlation module: receives the output results of the fault detection and positioning module, synchronizes the time data of each node in the distribution network, and uses event correlation technology to analyze the temporal and spatial relationship of fault events, thereby optimizing the output results of the fault detection and positioning module. The processed data is transmitted to the spatial information integration module;

[0087] Spatial information integration module: Receives the fault location information optimized by the synchronous association module, integrates it with the geographic information system (GIS) data, and generates a visual fault distribution map.

[0088] The environmental perception module includes a temperature monitoring unit, a humidity monitoring unit, a meteorological condition monitoring unit, and an electromagnetic interference monitoring unit; wherein:

[0089] Temperature monitoring unit: includes multiple temperature sensors, which are used to collect temperature data of different nodes in the distribution network in real time. The temperature sensor is based on the thermocouple principle and uses the electromotive force difference generated by metal at different temperatures to measure the temperature. The digital signal processor converts the analog signal into a digital signal and transmits it to the data fusion processing module in real time.

[0090] Humidity monitoring unit: includes multiple humidity sensors, which are used to collect humidity data of different nodes in the distribution network in real time. The humidity sensor is based on the principle of capacitive humidity measurement, and uses the change in capacitance caused by humidity change to measure humidity. The signal is amplified and filtered through the signal conditioning circuit, and then transmitted to the data fusion processing module;

[0091] Meteorological condition monitoring unit: It includes multiple meteorological sensors, which are used to collect meteorological data of the distribution network area in real time, including wind speed, wind direction, rainfall and atmospheric pressure. The meteorological sensors are based on mechanical and electronic measurement principles, and measure meteorological parameters through mechanical movement and electronic signal conversion, and transmit the collected data to the data fusion processing module;

[0092] Electromagnetic interference monitoring unit: includes multiple electromagnetic interference sensors, which are used to detect electromagnetic interference signals in the distribution network in real time. The electromagnetic interference sensor is based on the principle of electromagnetic induction. It uses the induced electromotive force generated by the induction coil when the electromagnetic field changes to detect the electromagnetic interference signal, and after processing through filtering and amplification circuits, the data is transmitted to the data fusion processing module; through the combination of the above units, the operating environment of the distribution network can be monitored more accurately in real time, including temperature, humidity, meteorological conditions and electromagnetic interference parameters, ensuring high precision and high real-time performance of the data, and providing a reliable foundation for subsequent data fusion processing and fault detection.

[0093] The data fusion processing module includes a data receiving unit, a preprocessing unit, an improved Kalman filter unit, a Bayesian network processing unit and a data output unit; wherein:

[0094] Data receiving unit: used to receive the operating environment data from the environment perception module and the electrical quantity data of each node in the distribution network. The data receiving unit is connected to the environment perception module and the sensor interface of each node in the distribution network through a high-speed data bus to ensure real-time and accurate data transmission;

[0095] Preprocessing unit: performs preliminary processing on the received multi-source data, including data format conversion, noise filtering and data alignment;

[0096] Improved Kalman filter unit: Dynamically filter and estimate the state of preprocessed data. The improved Kalman filter unit is based on the extended Kalman filter algorithm and combines the dynamic model of the distribution network to perform state prediction and error correction on the operating environment data and electrical quantity data to extract preliminary fault-related information;

[0097] Bayesian network processing unit: performs probability inference and data fusion on the preliminary information output by the Kalman filter unit. Specifically, it uses the conditional probability distribution model of the Bayesian network to fuse multi-source data, and combines historical fault data and prior knowledge to deeply extract effective information related to the fault.

[0098] Data output unit: Through the preset standard communication protocol and interface, the fused data is sent to the downstream module in real time to ensure the integrity and accuracy of the information; through the combination of the above units, multi-source data can be processed and fused efficiently and accurately. The algorithm combining the improved Kalman filter and Bayesian network improves the dynamic response and probabilistic inference capabilities of data processing, ensuring the accuracy and reliability of fault detection in complex environments. This multi-source data fusion processing technology provides a high-quality data foundation for the fault detection and positioning module, significantly improving the speed and accuracy of distribution network fault positioning, and ensuring the safe and stable operation of the power grid.

[0099] The improved Kalman filter unit includes:

[0100] State initialization: define the initial state variable x of the system 0 and the initial covariance matrix P 0 , used to describe the initial state and initial uncertainty of the system;

[0101] Prediction step: According to the dynamic model of the distribution network, the system state is predicted through the state transition equation, and the process noise covariance Q is used 1 Correct the prediction results; the specific calculation formula is:

[0102] x k|k-1 =f(x k-1 ,u k-1 );

[0103] Among them, x k|k-1 is the predicted state from time k-1 to time k; u k-1 is the control input at time k-1; F 1 is the Jacobian matrix of the state transfer matrix; Q 1 is the process noise covariance;

[0104] Measurement update step: Correct the predicted state according to the measurement equation, and correct the state by measuring the predicted value and the Kalman gain. The specific correction formula is:

[0105] z k|k-1 =h(x k|k-1 );

[0106]

[0107] Among them, z k|k-1 is the measurement prediction value vector, which represents the measurement value calculated according to the predicted state; H 1 is the Jacobian matrix of the measurement matrix, which represents the linear approximation of the measurement model; S 1 is the measurement prediction error covariance matrix, which represents the covariance of the measurement error; R 1 is the measurement noise covariance matrix, representing the measurement noise; K 1 is the Kalman gain matrix, which represents the weight of the prediction error correction;

[0108] State update: Use the Kalman gain to correct the predicted state to obtain the updated state estimate and covariance matrix. The specific formula is:

[0109] x k =x k|k-1 +K 1 (z k -z k|k-1 );

[0110] Among them, x k is the updated state estimation vector, which represents the corrected system state; P k is the updated covariance matrix, which represents the corrected state uncertainty; z k is the actual measurement value vector, which represents the actual data obtained from the environment perception module and the distribution network nodes;

[0111] Fault information extraction: Estimate x based on the updated state k and the covariance matrix P k , extract preliminary fault-related information; the above steps use the extended Kalman filter algorithm, combined with the dynamic model of the distribution network, to extract preliminary fault-related information, which significantly improves the accuracy and response speed of fault detection, provides a high-quality data foundation for the fault detection and positioning module, and ensures the safe and stable operation of the distribution network.

[0112] The Bayesian network processing unit includes:

[0113] Node definition: Define the nodes in the Bayesian network, where each node represents a random variable, specifically the state estimate x output by the Kalman filter unit k 、Historical fault data k and prior knowledge p k ;

[0114] Edge definition: Define the edges in the Bayesian network to represent the dependencies between nodes. Use the conditional probability distribution model to describe the dependencies. For example, the state estimate xk Reliance on historical failure data and prior knowledge;

[0115] Conditional probability table construction: Based on historical data and expert knowledge, a conditional probability table (CPT) is constructed to describe the probability distribution of each node given its parent node. k Given the historical fault data h k and prior knowledge p k The conditional probability in this case can be expressed as: P(x k =x|h k =h,p k =p);

[0116] Probabilistic inference: preliminary information x on the output of the Kalman filter unit k Probabilistic inference is performed, and the posterior probability of a fault is calculated by combining historical fault data and prior knowledge. The Bayesian inference formula is:

[0117] Among them, z k is the actual measurement value obtained from the environment perception module and the distribution network node; P(z k |x k ) is the likelihood function; P(x k ) is the prior probability; P(z k ) as evidence;

[0118] Data fusion: The preliminary information is fused using the Bayesian network, and the conditional probability distribution model and historical data are combined to deeply extract the fault-related information. The fused information includes the updated fault probability distribution and fault location estimation. The fusion processing formula is: Among them, F k is the fused fault information; is the i-th state estimate; is the posterior probability of the ith state estimate; through the working principles of the above steps, the system can effectively perform probabilistic inference and data fusion on the preliminary information output by the Kalman filter unit, and use the conditional probability distribution model of the Bayesian network, combined with historical fault data and prior knowledge, to deeply extract effective information related to the fault, thereby improving the accuracy and reliability of fault location, providing solid data support for subsequent fault handling, and ensuring the stable operation of the distribution network.

[0119] The fault detection and positioning module includes a signal preprocessing unit, a time-frequency analysis unit, a fault identification unit, a positioning unit and a result transmission unit; wherein:

[0120] Signal preprocessing unit: used to receive the effective information provided by the data fusion processing module and preprocess the signal, including denoising and normalization processing, to improve the signal quality and analysis accuracy. The preprocessed signal is transmitted to the time-frequency analysis unit;

[0121] Time-frequency analysis unit: Use improved wavelet transform to perform time-frequency analysis on the preprocessed signal. Through the preset mother wavelet function of the distribution network signal characteristics, the time domain signal is converted to the time-frequency domain, so that the fault characteristics are more obvious at different frequencies and time scales, thereby improving the resolution and recognition ability of the fault signal to capture the transient characteristics of the fault signal. The extracted fault feature signal is transmitted to the fault recognition unit;

[0122] Fault identification unit: Combines convolutional neural network to identify the fault type of the extracted fault feature signal. CNN is a deep learning algorithm that has the ability to automatically extract and learn signal features. Specifically, the fault feature signal is input into the trained convolutional neural network model. Through multi-layer convolution and pooling operations, high-level features are extracted and classified to identify the fault type. The identification results are transmitted to the positioning unit.

[0123] Positioning unit: Based on the fault type information output by the fault identification unit, the fault is initially located. By analyzing the time-frequency characteristics and spatial distribution information of the fault signal, the specific location of the fault is determined. The positioning information and fault type information are integrated to generate the fault detection result.

[0124] Result transmission unit: transmits the fault detection results and positioning information to the synchronization association module, specifically through standard communication protocols and data interfaces to ensure the integrity and real-time nature of the information, and support the synchronization association module for further analysis and optimization; by disclosing the working principle and unit composition of the fault detection and positioning module in detail, this system can efficiently and accurately detect and locate faults in the distribution network, use the improved wavelet transform for time-frequency analysis, extract accurate fault feature signals, and combine convolutional neural networks for fault type identification and positioning, thereby improving the accuracy and response speed of fault detection, providing high-quality data support for subsequent fault processing, and ensuring the stable operation of the distribution network.

[0125] The time-frequency analysis unit includes:

[0126] Select mother wavelet function: According to the characteristics of the distribution network fault signal, select a suitable mother wavelet function ψ(t), which includes Daubechies wavelet, Haar wavelet and Symlet wavelet. The selected mother wavelet function needs to have good time-frequency localization characteristics in order to effectively capture the transient characteristics of the fault signal.

[0127] Signal decomposition: The preprocessed time domain signal x(t) is decomposed into different frequencies and time scales through wavelet transform. The formula of continuous wavelet transform (CWT) is: Among them, W x (a, b) are wavelet coefficients, representing the signal component at scale a and position b; ψ is the mother wavelet function; a is the scale parameter; b is the translation parameter; ψ * is the complex conjugate of the mother wavelet function;

[0128] Multi-scale analysis: By adjusting the scale parameter a and the translation parameter b, the signal is analyzed at multiple scales. The small scale corresponds to high-frequency details, which is used to capture the transient characteristics of the signal; the large scale corresponds to low-frequency information, which is used to reflect the overall trend of the signal. The distribution of fault characteristics at different scales and times can be observed through the wavelet coefficient graph;

[0129] Feature extraction: By analyzing the wavelet coefficient graph, the fault feature signal is extracted. For transient fault signals, the feature will appear as a sudden change in the high-frequency component. The energy or amplitude of the wavelet coefficient is calculated to quantify the fault feature. Specifically, the wavelet energy E(a,b) at scale a and time b is calculated. The formula is: E(a,b)=|W x (a,b)| 2 ,The changes of the reaction wavelet energy at different scales can reveal the transient characteristics of the fault signal;

[0130] Characteristic signal reconstruction: According to the extracted characteristic signal, the wavelet coefficients are inversely transformed to reconstruct the fault characteristic signal. The reconstruction formula of discrete wavelet transform (DWT) is: x(t) = ∑ a,b W x (a,b)ψ a,b (t), where ψ a,b (t) is the mother wavelet function at scale a and position b, and the reconstructed fault feature signal is used for subsequent fault identification and location. Through the above steps, the system can efficiently and accurately perform time-frequency analysis on the preprocessed time domain signal, and use the preset mother wavelet function of the distribution network signal characteristics to convert the signal into the time-frequency domain, so that the fault characteristics are more obvious at different frequencies and time scales, thereby improving the resolution and recognition ability of the fault signal. This method effectively captures the transient characteristics of the fault signal, provides high-quality feature signals for subsequent fault identification and location, and significantly improves the accuracy and response speed of fault detection.

[0131] The positioning unit includes:

[0132] Receiving fault type information: The positioning unit first receives the fault type information from the fault identification unit, including the fault type and its corresponding signal feature description;

[0133] Characteristic signal reconstruction: Reconstruct the fault characteristic signal to obtain specific time-frequency characteristics and spatial distribution information;

[0134] Time-frequency analysis of fault signals: Analyze the time-frequency characteristics of the reconstructed signal, especially the transient characteristics, including spikes and breaks. This characteristic is more obvious in the high-frequency components of the wavelet transform. By quantitatively analyzing the intensity and distribution of the transient characteristics, the time point of the fault occurrence can be determined;

[0135] Spatial distribution analysis: Combine the topology of the distribution network and the known sensor locations to analyze the propagation and attenuation of fault characteristic signals at different locations. Use the speed and direction of signal propagation and the signal strength received from different sensors to calculate the potential location of the fault source.

[0136] Accurately locate the fault location: Based on the results of time-frequency feature analysis and spatial distribution analysis, the least squares estimation or triangulation positioning algorithm is used to accurately calculate the specific location of the fault. In this process, multiple data sources and fault indicators can be considered to improve the accuracy and reliability of positioning;

[0137] Integrate fault information and output: Integrate the positioning information and fault type information to generate fault detection results, including the fault type, the specific time and location of the fault, and the cause of the fault; through the above-mentioned steps of the public positioning unit, it is possible to effectively utilize the fault type information provided by the fault identification unit, and combine the time-frequency characteristics and spatial distribution information to perform preliminary fault positioning. By using characteristic signal reconstruction and advanced signal analysis technology, this system significantly improves the accuracy and efficiency of fault positioning. The ability to integrate fault information and output ensures the stability and safety of distribution network operation, and provides reliable data support for rapid response and fault repair.

[0138] The synchronization association module includes a time synchronization unit, an event extraction unit, a spatiotemporal association analysis unit, and a result optimization unit; wherein:

[0139] Time synchronization unit: used to receive the output results from the fault detection and positioning module, as well as the time data of each node in the distribution network, and achieve time synchronization through the preset network time protocol (NTP), ensuring that the timestamps of all nodes are consistent, eliminating clock drift and time errors, and providing an accurate time reference. The time synchronization formula is: Among them, T corrected is the corrected time, T received To receive the timestamp, T sent To send the timestamp, T reply is the reply timestamp, T transmit is the transmission timestamp;

[0140] Event extraction unit: extracts relevant event information from the output results of the fault detection and positioning module, including the time, type and preliminary positioning information of the fault. The event extraction unit analyzes the fault signal and identifies transient events related to the fault, providing basic data for subsequent spatiotemporal correlation analysis.

[0141] Time-space correlation analysis unit: Based on the synchronization time data provided by the time synchronization unit and the event information extracted by the event extraction unit, the event correlation technology is used to analyze the time-space relationship of the fault event. The specific analysis includes time series analysis.

[0142] Timing analysis uses synchronized time data to perform timing analysis on fault events to determine the sequence and time interval of events. By comparing the timestamps of different nodes, the fault propagation path and speed can be identified. The timing analysis formula is: Where Δt is the time difference between node i and node j, and is the timestamp of the corresponding node;

[0143] Spatial correlation analysis combines the geographic information system (GIS) data of the distribution network to analyze the spatial distribution and correlation of fault events. The geographical location of the event and the layout of sensors are used to determine the source of the fault and the scope of impact. The spatial correlation analysis formula is: Where d is the distance between two nodes, (x 1 ,y 1 ) and (x 2 ,y 2 ) is the position coordinate of the node;

[0144] Result optimization unit: According to the analysis results of the time-space correlation analysis unit, the preliminary positioning information of the fault detection and positioning module is optimized. Specifically, the preliminary positioning results are corrected by comprehensively analyzing the time sequence and spatial distribution of the fault events, and the positioning accuracy is improved. The optimized positioning information and fault type information are integrated to generate the final fault detection result. The optimization positioning formula is: Among them, L optimized is the optimized positioning result, L initial is the initial positioning result, L i is the positioning information of the ith node, α and β are weighting coefficients; through the combination of the above units, the time data of each node in the distribution network can be effectively synchronized, and the event correlation technology is used to analyze the spatiotemporal relationship of fault events, thereby optimizing the output results of the fault detection and positioning module. By using precise time synchronization and event correlation analysis technology, this system significantly improves the accuracy and reliability of fault location, provides strong data support for the rapid response and processing of distribution network faults, and ensures the stable operation of the power grid.

[0145] The spatial information integration module includes a data receiving unit, a data conversion unit, a spatial data processing unit and a visualization generation unit; wherein:

[0146] Data receiving unit: used to receive optimized fault location information from the synchronous association module and geographic data obtained from the geographic information system (GIS). The data receiving unit exchanges data with other modules through standard data interfaces and communication protocols to ensure the real-time and accuracy of data.

[0147] Data conversion unit: converts the received optimized fault location information into a format that can be recognized by the geographic information system, and uses the spatial reference system (SRS) and projection conversion method to match the location information with the geographic coordinates to ensure accurate mapping of the fault location information in the geographic space;

[0148] Spatial data processing unit: integrates the converted fault location information with the geographic information system data. The spatial data processing unit uses spatial analysis technology to process and analyze the fault information, including spatial overlay, buffer analysis and spatial interpolation operations, to generate accurate fault distribution information;

[0149] Visualization generation unit: Based on the integrated processed fault location information, a visualized fault distribution map is generated. The visualization generation unit uses the graphics rendering function of the GIS software to display the fault information in a graphical manner, including information on the fault point, impact range and fault type, to ensure that the generated distribution map has high resolution and a clear hierarchical structure, which is convenient for operation and maintenance personnel to conduct intuitive monitoring and analysis; through the combination of the above units, the fault location information optimized by the synchronization association module can be efficiently integrated with the geographic information system data to generate a visualized fault distribution map. By using spatial data processing and visualization technology, this system significantly improves the display effect and operability of fault information, and provides strong support for the rapid location and processing of distribution network faults. The generated visualized fault distribution map helps operation and maintenance personnel to intuitively understand the location and impact range of the fault, ensuring the stable operation of the power grid.

[0150] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. Rapid fault location system for distribution network, It is characterized in that It includes an environmental perception module, a data fusion processing module, a fault detection and positioning module, a synchronization association module and a spatial information integration module; among which: Environmental perception module: used to monitor the distribution network operating environment in real time, including temperature, humidity, meteorological conditions and electromagnetic interference parameters; Data fusion processing module: Receives the operating environment data transmitted by the environmental perception module, as well as the electrical quantity data of each node in the distribution network, processes the multi-source data through an algorithm combining improved Kalman filtering and Bayesian network, and extracts effective information related to the fault; Fault detection and positioning module: Based on the effective information provided by the data fusion processing module, the improved wavelet transform is first used to perform time-frequency analysis on the signal to extract the fault characteristic signal; then the convolutional neural network is combined to identify the fault type and preliminarily locate it, and finally the detection results and positioning information are transmitted to the synchronous association module; Synchronous correlation module: receives the output results of the fault detection and positioning module, synchronizes the time data of each node in the distribution network, and uses event correlation technology to analyze the temporal and spatial relationship of fault events, thereby optimizing the output results of the fault detection and positioning module; Spatial information integration module: receives the fault location information optimized by the synchronous association module, integrates it with the geographic information system data, and generates a visual fault distribution map.

2. The distribution network fault rapid location system according to claim 1, It is characterized in that The environment sensing module includes a temperature monitoring unit, a humidity monitoring unit, a meteorological condition monitoring unit and an electromagnetic interference monitoring unit; wherein: Temperature monitoring unit: includes multiple temperature sensors, which are used to collect temperature data of different nodes in the distribution network in real time. The temperature sensor is based on the thermocouple principle and uses the electromotive force difference generated by metal at different temperatures to measure the temperature. The digital signal processor converts the analog signal into a digital signal and transmits it to the data fusion processing module in real time. Humidity monitoring unit: includes multiple humidity sensors, which are used to collect humidity data of different nodes of the distribution network in real time. The humidity sensor is based on the principle of capacitance humidity measurement, and uses the change of capacitance value caused by humidity change to measure humidity. The signal is amplified and filtered through the signal conditioning circuit, and then transmitted to the data fusion processing module; Meteorological condition monitoring unit: including multiple meteorological sensors for real-time collection of meteorological data in the distribution network area, including wind speed, wind direction, rainfall and atmospheric pressure; the meteorological sensors are based on mechanical and electronic measurement principles, measure meteorological parameters through mechanical movement and electronic signal conversion, and transmit the collected data to the data fusion processing module; Electromagnetic interference monitoring unit: includes multiple electromagnetic interference sensors, which are used to detect electromagnetic interference signals in the distribution network in real time. The electromagnetic interference sensor is based on the principle of electromagnetic induction and uses the induced electromotive force generated by the induction coil when the electromagnetic field changes to detect the electromagnetic interference signal. After being processed by filtering and amplifying circuits, the data is transmitted to the data fusion processing module.

3. The distribution network fault rapid location system according to claim 1, It is characterized in that The data fusion processing module includes a data receiving unit, a preprocessing unit, an improved Kalman filter unit, a Bayesian network processing unit and a data output unit; wherein: Data receiving unit: used to receive the operating environment data from the environment perception module and the electrical quantity data of each node in the distribution network. The data receiving unit is connected to the environment perception module and the sensor interface of each node in the distribution network through a high-speed data bus to ensure real-time and accurate data transmission; Preprocessing unit: performs preliminary processing on the received multi-source data, including data format conversion, noise filtering and data alignment; Improved Kalman filter unit: Dynamically filter and estimate the state of preprocessed data. The improved Kalman filter unit is based on the extended Kalman filter algorithm and combines the dynamic model of the distribution network to perform state prediction and error correction on the operating environment data and electrical quantity data to extract preliminary fault-related information; Bayesian network processing unit: performs probability inference and data fusion on the preliminary information output by the Kalman filter unit. Specifically, it uses the conditional probability distribution model of the Bayesian network to fuse multi-source data, and combines historical fault data and prior knowledge to deeply extract effective information related to the fault. Data output unit: Send the fused data to the downstream modules in real time through the preset standard communication protocol and interface.

4. The distribution network fault rapid location system according to claim 3, It is characterized in that The improved Kalman filter unit comprises: State initialization: define the initial state variable x of the system 0 and the initial covariance matrix P 0 , used to describe the initial state and initial uncertainty of the system; Prediction step: According to the dynamic model of the distribution network, the system state is predicted through the state transition equation, and the process noise covariance Q is used 1 Correct the prediction results; the specific calculation formula is: x k|k-1 =f(x k-1 ,u k-1 ); Among them, x k|k-1 is the predicted state from time k-1 to time k; u k-1 is the control input at time k-1; F 1 is the Jacobian matrix of the state transfer matrix; Q 1 is the process noise covariance; Measurement update step: Correct the predicted state according to the measurement equation, and correct the state by measuring the predicted value and the Kalman gain. The specific correction formula is: z k|k-1 =h(x k|k-1 ); Among them, z k|k-1 is the measurement prediction value vector, which represents the measurement value calculated according to the predicted state; H 1 is the Jacobian matrix of the measurement matrix, which represents the linear approximation of the measurement model; S 1 is the measurement prediction error covariance matrix, which represents the covariance of the measurement error; R 1 is the measurement noise covariance matrix, representing the measurement noise; K 1 is the Kalman gain matrix, which represents the weight of the prediction error correction; State update: Use the Kalman gain to correct the predicted state to obtain the updated state estimate and covariance matrix. The specific formula is: x k =x k|k-1 +K 1 (z k -z k|k-1 ); Among them, x k is the updated state estimation vector, which represents the corrected system state; P k is the updated covariance matrix, which represents the corrected state uncertainty; z k is the actual measurement value vector, which represents the actual data obtained from the environment perception module and the distribution network nodes; Fault information extraction: Estimate x based on the updated state k and the covariance matrix P k , extract preliminary fault-related information.

5. The distribution network fault rapid location system according to claim 4, It is characterized in that The Bayesian network processing unit comprises: Node definition: Define the nodes in the Bayesian network, where each node represents a random variable, specifically the state estimate x output by the Kalman filter unit k 、Historical fault data k and prior knowledge p k ; Edge definition: Define the edges in the Bayesian network to represent the dependency between nodes, and use the conditional probability distribution model to describe the dependency; Conditional probability table construction: Based on historical data and expert knowledge, a conditional probability table is constructed to describe the probability distribution of each node under a given parent node. k Given the historical fault data h k and prior knowledge p k The conditional probability in this case can be expressed as: P(x k =x|h k =h,p k =p); Probabilistic inference: preliminary information x on the output of the Kalman filter unit k Probabilistic inference is performed, and the posterior probability of a fault is calculated by combining historical fault data and prior knowledge. The Bayesian inference formula is: Among them, z k is the actual measurement value obtained from the environment perception module and the distribution network node; P(z k |x k ) is the likelihood function; P(x k ) is the prior probability; P(z k ) as evidence; Data fusion: The preliminary information is fused using the Bayesian network, and the conditional probability distribution model and historical data are combined to deeply extract the fault-related information. The fused information includes the updated fault probability distribution and fault location estimation. The fusion processing formula is: Among them, F k is the fused fault information; is the i-th state estimate; is the estimated posterior probability of the i-th state.

6. The distribution network fault rapid location system according to claim 1, It is characterized in that The fault detection and positioning module includes a signal preprocessing unit, a time-frequency analysis unit, a fault identification unit, a positioning unit and a result transmission unit; wherein: Signal preprocessing unit: used to receive the effective information provided by the data fusion processing module and preprocess the signal, including denoising and normalization, to improve signal quality and analysis accuracy; Time-frequency analysis unit: Use improved wavelet transform to perform time-frequency analysis on the pre-processed signal. Through the preset mother wavelet function of the distribution network signal characteristics, the time domain signal is converted to the time-frequency domain, making the fault characteristics more obvious at different frequencies and time scales, thereby improving the resolution and recognition ability of the fault signal to capture the transient characteristics of the fault signal; Fault identification unit: Combines convolutional neural network to identify the fault type of the extracted fault feature signal. Specifically, the fault feature signal is input into the trained convolutional neural network model, and through multi-layer convolution and pooling operations, high-level features are extracted and classified to identify the fault type. Positioning unit: Based on the fault type information output by the fault identification unit, the fault is initially located. By analyzing the time-frequency characteristics and spatial distribution information of the fault signal, the specific location of the fault is determined. The positioning information and fault type information are integrated to generate the fault detection result. Result transmission unit: transmits fault detection results and location information to the synchronization association module.

7. The distribution network fault rapid location system according to claim 6, It is characterized in that The time-frequency analysis unit comprises: Select mother wavelet function: According to the characteristics of distribution network fault signal, select a suitable mother wavelet function ψ(t), which includes Daubechies wavelet, Haar wavelet and Symlet wavelet; Signal decomposition: The preprocessed time domain signal x(t) is decomposed into different frequencies and time scales through wavelet transform. The formula of continuous wavelet transform is: Among them, W x (a, b) are wavelet coefficients, representing the signal component at scale a and position b; ψ is the mother wavelet function; a is the scale parameter; b is the translation parameter; ψ * is the complex conjugate of the mother wavelet function; Multi-scale analysis: By adjusting the scale parameter a and the translation parameter b, the signal is analyzed at multiple scales. The small scale corresponds to high-frequency details, which is used to capture the transient characteristics of the signal; the large scale corresponds to low-frequency information, which is used to reflect the overall trend of the signal. The distribution of fault characteristics at different scales and times can be observed through the wavelet coefficient graph; Feature extraction: By analyzing the wavelet coefficient graph, the fault feature signal is extracted. For transient fault signals, the feature will appear as a sudden change in the high-frequency component. The energy or amplitude of the wavelet coefficient is calculated to quantify the fault feature. Characteristic signal reconstruction: According to the extracted characteristic signal, the wavelet coefficients are inversely transformed to reconstruct the fault characteristic signal. The reconstruction formula of discrete wavelet transform is: x(t) = ∑ a,b W x (a,b)ψ a,b (t), where ψ a,b (t) is the mother wavelet function at scale a and position b.

8. The distribution network fault rapid location system according to claim 7, It is characterized in that The positioning unit comprises: Receiving fault type information: The positioning unit first receives the fault type information from the fault identification unit, including the fault type and its corresponding signal feature description; Characteristic signal reconstruction: Reconstruct the fault characteristic signal to obtain specific time-frequency characteristics and spatial distribution information; Time-frequency analysis of fault signals: Analyze the time-frequency characteristics of the reconstructed signal, especially the transient characteristics, including spikes and breaks. This characteristic is more obvious in the high-frequency components of the wavelet transform. By quantitatively analyzing the intensity and distribution of the transient characteristics, the time point of the fault occurrence can be determined; Spatial distribution analysis: Combine the topology of the distribution network and the known sensor locations to analyze the propagation and attenuation of fault characteristic signals at different locations. Use the speed and direction of signal propagation and the signal strength received from different sensors to calculate the potential location of the fault source. Accurately locate the fault location: Based on the results of time-frequency feature analysis and spatial distribution analysis, the least squares estimation or triangulation positioning algorithm is used to accurately calculate the specific location of the fault; Integrate fault information and output: Integrate location information and fault type information to generate fault detection results, including fault type, specific time and location of the fault, and cause of the fault.

9. The distribution network fault rapid location system according to claim 1, It is characterized in that The synchronization association module includes a time synchronization unit, an event extraction unit, a spatiotemporal association analysis unit, and a result optimization unit; wherein: Time synchronization unit: used to receive the output results from the fault detection and positioning module, as well as the time data of each node in the distribution network, and achieve time synchronization through the preset network time protocol to ensure that the timestamps of all nodes are consistent, eliminate clock drift and time error, and provide an accurate time reference; Event extraction unit: extracts relevant event information from the output results of the fault detection and positioning module, including the time, type and preliminary positioning information of the fault. The event extraction unit analyzes the fault signal and identifies transient events related to the fault, providing basic data for subsequent spatiotemporal correlation analysis. Time-space correlation analysis unit: Based on the synchronization time data provided by the time synchronization unit and the event information extracted by the event extraction unit, the event correlation technology is used to analyze the time-space relationship of the fault event. The specific analysis includes time series analysis. The timing analysis uses the synchronous time data to perform timing analysis on the fault events, determine the sequence and time interval of the events, and identify the fault propagation path and speed by comparing the timestamps of different nodes. The timing analysis formula is: Where Δt is the time difference between node i and node j, and is the timestamp of the corresponding node; The spatial correlation analysis combines the geographic information system data of the distribution network to analyze the spatial distribution and correlation of fault events, and uses the geographical location of the event and the sensor layout to determine the source of the fault and the scope of influence. The spatial correlation analysis formula is: Where d is the distance between two nodes, (x 1 ,y 1 ) and (x 2 ,y 2 ) is the position coordinate of the node; Result optimization unit: According to the analysis results of the time-space correlation analysis unit, the preliminary positioning information of the fault detection and positioning module is optimized. Specifically, the preliminary positioning results are corrected by comprehensively analyzing the time sequence and spatial distribution of the fault events, and the positioning accuracy is improved. The optimized positioning information and fault type information are integrated to generate the final fault detection result. The optimization positioning formula is: Among them, L optimized is the optimized positioning result, L initial is the initial positioning result, L i is the location information of the i-th node, α and β are weighting coefficients.

10. The distribution network fault rapid location system according to claim 1, It is characterized in that The spatial information integration module includes a data receiving unit, a data conversion unit, a spatial data processing unit and a visualization generation unit; wherein: Data receiving unit: used to receive optimized fault location information from the synchronization association module and geographic data obtained from the geographic information system; Data conversion unit: converts the received optimized fault location information into a format that can be recognized by the geographic information system, and uses the spatial reference system and projection conversion method to match the location information with the geographic coordinates to ensure accurate mapping of the fault location information in the geographic space; Spatial data processing unit: integrates the converted fault location information with the geographic information system data. The spatial data processing unit uses spatial analysis technology to process and analyze the fault information, including spatial overlay, buffer analysis and spatial interpolation operations, to generate accurate fault distribution information; Visualization generation unit: Based on the integrated processed fault location information, a visual fault distribution map is generated. The visualization generation unit uses the graphics rendering function of the GIS software to display the fault information in a graphical manner, including information on the fault point, impact range and fault type.

Citation Information

Patent Citations

  • Improved agricultural operation area measuring terminal and method

    CN109029243A

  • Fault prediction method for power distribution network

    CN114139753A

  • Smart city heterogeneous sensor multi-target cooperative positioning method and storage medium thereof

    CN114659517A

  • Aviation obstruction beacon remote monitoring method and system

    CN117972580A

  • Device for monitoring the integrity of information delivered by a hybrid INS / GNSS system

    EP1714166A1

Cited By

  • System and method for positioning grounding fault of power distribution network

    CN119667389A

  • Method and equipment for synchronously measuring electrical impedance and magnetic flux density of current collection circuit

    CN119689178A

  • Traveling wave signal-based coiled cable fault detection method and system

    CN119738662A

  • Method and system for verifying sensor data of injection molding machine and judging data normality

    CN119740165A

  • Fault detection and root cause analysis system for multi-source energy system

    CN119760607A