CN-122017470-A - Power distribution network insulation fault detection and positioning method and system based on multi-source transient data

CN122017470ACN 122017470 ACN122017470 ACN 122017470ACN-122017470-A

Abstract

The invention relates to the technical field of relay protection and fault detection of power systems, and discloses a method and a system for detecting and positioning insulation faults of a power distribution network based on multi-source transient data. According to the method and the system for detecting and positioning the insulation faults of the power distribution network based on the multi-source transient data, the fault branch is directly identified through the unique mapping relation between the high-frequency attenuation spectrum of the traveling wave head and the path topology of the power distribution network, so that the unique positioning result is realized, no pseudo point interference is caused, the change of the network operation mode can be automatically adapted through the dynamic topology image and the branch fingerprint library, strong robustness and practical value are shown for weak fault signals such as high-resistance grounding and the like and complex network structures, and the reliability and operation and maintenance efficiency of the fault positioning of the power distribution network are remarkably improved.

Inventors

WANG MINZHEN
ZHANG GUANGXIN
Yue keyu
LU XIAOHONG
SHEN KEJU
PEI YUJIE
HAN YUXIAN
ZHENG YU
HE ZHENGJIE

Assignees

宁波大学科学技术学院
国网吉林省电力有限公司辽源供电公司
国网浙江省电力有限公司慈溪市供电公司
国网辽宁省电力有限公司抚顺供电公司
长春晟德科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. The method for detecting and positioning the insulation faults of the power distribution network based on the multi-source transient data is characterized by comprising the following steps of: S1, synchronously collecting transient traveling wave signals generated by faults at monitoring points of a power distribution network, and capturing fault initial traveling wave head data reaching each monitoring point; S2, performing time-frequency analysis on the captured initial traveling wave head data, extracting energy values of wave head signals in at least two preset characteristic high-frequency bands and energy values in a reference low-frequency band; S3, carrying out modal decomposition on the initial traveling wave head data to obtain a linear mode component and a zero mode component, and independently executing a step S2 on wave heads of the linear mode component and the zero mode component to obtain a linear mode high-frequency attenuation spectrum characteristic vector and a zero mode high-frequency attenuation spectrum characteristic vector, so as to form a multi-modal verification characteristic pair; s4, generating a dynamic topology according to the real-time switching state of the power distribution network, and generating a standard branch attenuation fingerprint in advance for each potential fault branch based on the dynamic topology, wherein the standard branch attenuation fingerprint defines a theoretical high-frequency attenuation spectrum feature vector set which is transmitted to each monitoring point when a fault occurs in a corresponding branch; s5, collecting actually-measured multi-mode verification feature pairs of all monitoring points to form a multi-source feature set, carrying out matching degree calculation on the multi-source feature set and theoretical feature sets of candidate branches in a fingerprint library, and determining a fault branch according to a matching degree calculation result.
2. The method for detecting and locating an insulation fault of a power distribution network based on multi-source transient data of claim 1, wherein in step S2, the construction process of the high-frequency attenuation spectrum feature vector is that the energy of the reference low frequency band is set as First, the The energy of the characteristic high frequency band is The corresponding attenuation coefficient The calculation formula of (2) is as follows: The high-frequency attenuation spectrum characteristic vector of the monitoring point is 。
3. The method for detecting and positioning insulation faults of a power distribution network based on multi-source transient data as claimed in claim 1, wherein in step S5, the matching degree calculation adopts a weighted multi-mode distance measure to calculate the multi-source feature set And candidate branch Theoretical fingerprint of (2) Overall distance between The formula of (2) is: wherein And The mahalanobis distance or euclidean distance between the linear and zero mode feature vectors and the theoretical value, As the weight coefficient of the light-emitting diode, Matching degree score The calculation formula of (2) is 。
4. The method for detecting and positioning insulation faults of a power distribution network based on multi-source transient data according to claim 1, wherein a combined criterion based on wavelet transformation detail coefficient energy mutation rate and instantaneous amplitude of high-frequency components is adopted to capture an initial traveling wave head, and a time window is set to isolate subsequent reflected waves.
5. The method for detecting and positioning the insulation fault of the power distribution network based on the multi-source transient data, which is disclosed in claim 1, is characterized in that in the step S2, S conversion or wavelet conversion is adopted for the time-frequency analysis, and the characteristic high-frequency band is located in a frequency range from 150kHz to 1 MHz.
6. The method for detecting and positioning insulation faults of a power distribution network based on multi-source transient data according to claim 1, wherein the standard branch attenuation fingerprint is generated through electromagnetic transient simulation or a machine learning model trained based on historical data and simulation data.
7. A multisource transient data based distribution network insulation fault detection and localization system for implementing the method of any one of claims 1-6, comprising: The distributed monitoring terminals are deployed at key nodes of the power distribution network and are used for synchronously collecting transient traveling wave signals and completing initial wave head capturing and local feature extraction; The master station analysis system is connected with all the monitoring terminals through a communication network and is used for realizing dynamic topology management, construction and updating of a branch attenuation fingerprint library, multi-source feature matching calculation and fault branch identification, wherein the monitoring terminals at least comprise a broadband high-speed wave recording module, an initial wave head triggering module and a local feature extraction module, and the master station analysis system at least comprises a dynamic topology management module, a branch fingerprint library management module and a fault identification analysis module.
8. The system for detecting and positioning insulation faults of a power distribution network based on multi-source transient data of claim 7, wherein the local feature extraction module is used for executing high-frequency attenuation spectrum feature vector extraction and multi-mode coupling feature generation, and the fault identification analysis module is used for executing multi-source feature matching calculation and outputting fault branch positioning results and confidence.
9. The system for detecting and locating insulation faults of a power distribution network based on multi-source transient data of claim 7 further comprising a high speed communication network for enabling reliable transmission of time synchronization signals and data information between the monitoring terminal and the primary station analysis system.
10. The system for detecting and positioning the insulation faults of the power distribution network based on the multi-source transient data of claim 9, wherein the high-speed communication network is a hybrid communication architecture adopting an optical fiber private network and a 5G wireless network, and the system is characterized in that time synchronization between monitoring terminals and data transmission between the monitoring terminals and a master station analysis system are ensured.

Description

Power distribution network insulation fault detection and positioning method and system based on multi-source transient data Technical Field The invention relates to the technical field of relay protection and fault detection of power systems, in particular to a method and a system for detecting and positioning insulation faults of a power distribution network based on multi-source transient data. Background Traveling wave positioning is an important technical direction of power distribution network fault positioning, and the fault distance is calculated by measuring the time difference of high-frequency transient traveling wave signals generated by faults to reach different monitoring points. However, in a tree-shaped power distribution network with a complex structure, a complex refraction and reflection phenomenon occurs at impedance discontinuities such as branch points and tail ends of the traveling wave, so that a monitoring device may capture reflected waves from a non-fault path. When the arrival time of the reflected wave is confused with the direct wave time from the true fault point, the calculation based on the time difference may produce erroneous distance information, thereby forming a "false fault point" at the non-fault location. This problem is a long-term bottleneck that restricts the practical use of the traveling wave positioning technique. The prior art mainly improves the wave head detection precision, optimizes time synchronization or multi-terminal data fusion, but does not deviate from the traditional paradigm of measuring time, calculating distance and mapping position, and is essentially limited by the interference of reflected waves. Some approaches employing artificial intelligence attempt to bypass accurate modeling, but their "black box" nature results in poor interpretability and difficulty in accommodating topology changes. Therefore, a new positioning principle and method that can fundamentally immunize against the interference of traveling wave refraction and reflection are urgently needed. Disclosure of Invention The invention aims to solve the technical problems that in the prior art, the fault point is generated by the fact that traveling wave arrival time is dependent and reflected wave interference is easy to occur, and therefore, the method and the system for detecting and positioning the insulation fault of the power distribution network based on multi-source transient data are provided. In order to achieve the purpose, the application adopts the following technical scheme that the method for detecting and positioning the insulation faults of the power distribution network based on the multi-source transient data comprises the following steps: S1, synchronously collecting transient traveling wave signals generated by faults at monitoring points of a power distribution network, and capturing fault initial traveling wave head data reaching each monitoring point; S2, performing time-frequency analysis on the captured initial traveling wave head data, extracting energy values of wave head signals in at least two preset characteristic high-frequency bands and energy values in a reference low-frequency band; S3, carrying out modal decomposition on the initial traveling wave head data to obtain a line mode component and a zero mode component, and independently executing a step S2 on wave heads of the line mode component and the zero mode component to obtain a line mode high-frequency attenuation spectrum characteristic vector and a zero mode high-frequency attenuation spectrum characteristic vector to form a multi-modal verification characteristic pair; S4, generating a dynamic topology according to the real-time switching state of the power distribution network, and generating a standard branch attenuation fingerprint in advance for each potential fault branch based on the dynamic topology, wherein the standard branch attenuation fingerprint defines a theoretical high-frequency attenuation spectrum feature vector set which is transmitted to each monitoring point when a fault occurs in a corresponding branch; s5, collecting actually-measured multi-mode verification feature pairs of all monitoring points to form a multi-source feature set, carrying out matching degree calculation on the multi-source feature set and theoretical feature sets of candidate branches in a fingerprint library, and determining a fault branch according to a matching degree calculation result. Preferably, in step S2, the construction process of the high-frequency attenuation spectrum characteristic vector includes setting the energy of the reference low frequency band asFirst, theThe energy of the characteristic high frequency band isThe corresponding attenuation coefficientThe calculation formula of (2) is as follows: The high-frequency attenuation spectrum characteristic vector of the monitoring point is 。 Preferably, in step S5, the matching degree calculation uses a weighted multi-modal distance