CN-121145032-B - Secondary equipment fault diagnosis and positioning method and system based on association analysis

CN121145032BCN 121145032 BCN121145032 BCN 121145032BCN-121145032-B

Abstract

The application relates to the technical field of power automation, and discloses a secondary equipment fault diagnosis and positioning method and system based on correlation analysis. The method comprises the steps of calculating association weights by constructing a dynamic space-time association graph and adopting a double-exponential decay model based on topological weights, alarm time differences, system states and historical feedback. And through the collection of the multi-time scale alarm sequence and a graph propagation algorithm, the fault source probability of each sub-graph is calculated in an iterative mode, and the accurate fault source positioning result is output by combining the weighted fusion of the window credibility. By introducing a multi-level association and space-time weighting mechanism, the application not only enables fault diagnosis and positioning to be based on the current fault information, but also can fully utilize historical data feedback to adjust, thereby being better suitable for the dynamic operation environment and complex fault scene of the power system.

Inventors

YAN YITIAN
TANG XIAOJIE
SHEN QIHAO
YAN JIANGBO
HUANG YUNBING
ZHOU XUDONG
LV TAO
ZHOU NAN
XI QIANGYONG
DAI XUYI
YANG SHUO
WANG JIAN
XIA HAILIANG
SHEN TAO

Assignees

宁波送变电建设有限公司

Dates

Publication Date: 20260508
Application Date: 20251119

Claims (9)

1. A secondary equipment fault diagnosis and location method based on association analysis, the method comprising: Step S1, acquiring a topological connection relation of secondary equipment, calculating a dynamic space-time association weight through a double-exponential decay model according to a topological weight, an alarm time difference, a system running state and a historical feedback factor, and constructing a dynamic space-time association diagram, wherein the method comprises the steps of acquiring physical connection and logic association relation among a protection device, a measurement and control device, communication equipment and a telemechanical device in a power system, establishing a topological diagram of the secondary equipment comprising an equipment node set and a connection edge set, and respectively endowing the topological weight corresponding to the connection edge according to direct electrical connection, communication connection and logic association; calculating a time attenuation function through a double-exponential attenuation model based on an alarm time difference, wherein the time attenuation function comprises a fast component and a slow component, the time constant of the fast component is set to be 50 milliseconds, the time constant of the slow component is set to be 2 seconds, calculating a state modulation function according to a topology state vector, a load level vector and an equipment maintenance state vector in the current system running state, wherein the topology state vector records the switch positions of all circuit breakers, the load level vector records the load rates of all feeder lines, the equipment maintenance state vector records the running years and the historical failure times of equipment, and a historical feedback factor calculates according to the accurate contribution rate of a connecting edge in the past failure diagnosis; Step S2, setting a fast window, a medium speed window and a slow speed window, respectively collecting alarm information, and generating a corresponding fast alarm sequence, a medium speed alarm sequence and a slow alarm sequence; Step S3, mapping the rapid alarm sequence, the medium-speed alarm sequence and the slow alarm sequence to equipment nodes of the dynamic space-time correlation graph, and calculating the correlation strength among the nodes based on the dynamic space-time correlation weight to obtain a rapid correlation sub-graph, a medium-speed correlation sub-graph and a slow correlation sub-graph; S4, carrying out probability iterative computation on the fast associated subgraph, the medium-speed associated subgraph and the slow associated subgraph based on a graph propagation algorithm introducing attention weights to respectively obtain fast fault source probability distribution, medium-speed fault source probability distribution and slow fault source probability distribution; And S5, calculating the window credibility according to the alarm quantity and alarm level of each window, carrying out weighted fusion on the probability distribution of the fast fault source, the probability distribution of the medium-speed fault source and the probability distribution of the slow fault source, and outputting a fault source positioning result.
2. The correlation analysis-based secondary device fault diagnosis and localization method according to claim 1, wherein the step S2 comprises: setting the rapid window, collecting a protection device action signal, a switch position change signal and a fault wave recording starting signal through the rapid window, and generating the rapid alarm sequence according to the sequence of time stamps; Setting the medium speed window, collecting a telemetry out-of-limit alarm, a telemetry shift signal and a device self-checking abnormal signal through the medium speed window, and sequencing according to a time stamp to generate the medium speed alarm sequence; Setting the slow window, collecting communication delay alarm, equipment temperature alarm and memory alarm through the slow window, and generating the slow alarm sequence according to the time stamp ordering; And uniformly formatting the alarm information in the rapid alarm sequence, the medium-speed alarm sequence and the slow alarm sequence into alarm data tuples comprising equipment identification, alarm types, alarm moments and alarm levels, and respectively giving emergency, important and generally corresponding alarm weight coefficients according to the alarm levels.
3. The correlation analysis-based secondary device fault diagnosis and localization method according to claim 1, wherein the step S3 comprises: Respectively mapping alarm data elements in the rapid alarm sequence, the medium-speed alarm sequence and the slow alarm sequence to corresponding equipment nodes in the dynamic space-time correlation diagram; calculating the association strength between any two equipment nodes in the rapid alarm sequence based on the dynamic space-time association weight and the alarm weight coefficient, wherein the association strength is obtained by multiplying the dynamic space-time association weight by the alarm weight coefficient of the two nodes and the alarm type correlation coefficient; And carrying out sparsification treatment on the correlation intensity matrixes respectively calculated by the rapid alarm sequence, the medium-speed alarm sequence and the slow alarm sequence, and reserving node connecting edges with correlation intensity larger than a preset threshold value to respectively form the rapid correlation subgraph, the medium-speed correlation subgraph and the slow correlation subgraph.
4. The correlation analysis-based secondary device fault diagnosis and localization method according to claim 1, wherein the step S4 comprises: calculating the initial fault source probability of each equipment node according to the alarm weight coefficients of each equipment node in the fast associated subgraph, the medium-speed associated subgraph and the slow associated subgraph; Calculating probability propagation values of neighbor nodes to the current node based on the sum of the association strength and the node output weight among the equipment nodes in the rapid association subgraph, and updating the fault source probability of the current node according to the retention coefficient and the probability propagation values; Calculating the alarm feature similarity among nodes through the alarm type, the alarm time and the equipment type, calculating the attention weight based on the association strength and the alarm feature similarity, introducing the attention weight into the calculation of the probability propagation value, and performing multi-round iterative updating on the rapid association subgraph to obtain the rapid fault source probability distribution; And respectively processing the medium-speed associated subgraph and the slow-speed associated subgraph by adopting the same probability iterative computation mode to obtain the medium-speed fault source probability distribution and the slow-speed fault source probability distribution, and selecting the equipment node with the highest probability value from the probability distribution to form a corresponding candidate fault source node set.
5. The method for diagnosing and locating a fault of a secondary device based on association analysis according to claim 4, wherein the calculating a probability propagation value of a neighboring node to a current node based on a sum of association strength and node-out-degree weights among device nodes in the fast association subgraph, and updating a fault source probability of the current node according to a retention coefficient and the probability propagation value comprises: calculating the sum of the outbound weights of all the neighbor nodes, wherein the sum of the outbound weights is the sum of the association strengths of all the neighbor nodes pointed to by the neighbor nodes; Multiplying the current fault source probability of each neighbor node by the association strength of the current node, dividing the current fault source probability by the outbound weight sum of the neighbor node to obtain a single-node probability propagation value of each neighbor node to the current node, and summing the single-node probability propagation values of all neighbor nodes to obtain a total probability propagation value of the current node; Multiplying the initial fault source probability of the current node by a retention coefficient, multiplying the total probability propagation value by the complement of the retention coefficient, and adding the two to obtain the updated fault source probability of the current node.
6. The correlation analysis-based secondary device fault diagnosis and localization method according to claim 1, wherein the step S5 comprises: According to the alarm quantity and alarm level in the fast window, the medium speed window and the slow speed window, respectively calculating the reliability of the fast window, the reliability of the medium speed window and the reliability of the slow speed window, and carrying out normalization processing on the reliability of the three windows; based on the fast window credibility, the medium speed window credibility and the slow speed window credibility, carrying out weighted fusion on the fast fault source probability distribution, the medium speed fault source probability distribution and the slow fault source probability distribution to obtain global fault source probability; Correcting the global fault source probability according to the accuracy of taking the equipment node as a real fault source in the historical fault diagnosis record to obtain a final fault source probability; and sequencing the final fault source probability from high to low according to the numerical value, and outputting the fault source positioning result.
7. A correlation analysis-based secondary device fault diagnosis and localization system for implementing the correlation analysis-based secondary device fault diagnosis and localization method as set forth in any one of claims 1 to 6, the correlation analysis-based secondary device fault diagnosis and localization system comprising: The acquisition module is used for acquiring the topological connection relation of the secondary equipment, calculating the dynamic space-time association weight through a double-exponential decay model according to the topological weight, the alarm time difference, the system running state and the historical feedback factor, and constructing a dynamic space-time association diagram; The acquisition module is used for setting a fast window, a medium speed window and a slow speed window, respectively acquiring alarm information and generating a corresponding fast alarm sequence, a medium speed alarm sequence and a slow alarm sequence; The calculation module is used for mapping the rapid alarm sequence, the medium-speed alarm sequence and the slow alarm sequence to equipment nodes of the dynamic space-time correlation graph, and calculating the correlation strength among the nodes based on the dynamic space-time correlation weight to obtain a rapid correlation subgraph, a medium-speed correlation subgraph and a slow correlation subgraph; The iteration module is used for carrying out probability iterative computation on the fast associated subgraph, the medium-speed associated subgraph and the slow associated subgraph based on a graph propagation algorithm introducing attention weights to respectively obtain fast fault source probability distribution, medium-speed fault source probability distribution and slow fault source probability distribution; and the positioning module is used for calculating the window credibility according to the alarm quantity and alarm level of each window, carrying out weighted fusion on the probability distribution of the fast fault source, the probability distribution of the medium-speed fault source and the probability distribution of the slow fault source, and outputting a fault source positioning result.
8. A secondary device fault diagnosis and localization apparatus based on correlation analysis, characterized in that it comprises a memory and a processor, the memory storing a computer program executable on the processor, the processor implementing the secondary device fault diagnosis and localization method based on correlation analysis of any one of claims 1 to 6 when executing the computer program.
9. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, causes the processor to perform the correlation analysis-based secondary device fault diagnosis and localization method as claimed in any one of claims 1 to 6.

Description

Secondary equipment fault diagnosis and positioning method and system based on association analysis Technical Field The application relates to the technical field of power automation, in particular to a secondary equipment fault diagnosis and positioning method and system based on correlation analysis. Background In modern power systems, secondary devices (such as protection devices, measurement and control devices, communication devices, and telemechanical devices) play a critical role in ensuring safe and stable operation of the power system. The equipment is not only responsible for monitoring the running state of the power equipment in real time, but also can send out alarm signals in time when faults occur, and assists operators in fault diagnosis and treatment. Currently, secondary equipment fault diagnosis methods for power systems typically rely on manual analysis based on equipment alarms, or fault localization by simple rules. However, these conventional diagnostic methods have insufficient coping capability for complex fault scenarios, and are susceptible to false alarms and missed alarms, resulting in a great compromise in accuracy and timeliness of fault localization. The existing fault diagnosis method generally depends on single type of alarm information, ignores possible space-time correlation among alarms, and causes larger uncertainty in the positioning process of fault sources, and particularly in a complex system or a large-scale power network, the problems of diversity and timeliness of the alarm information are more remarkable. Because of inconsistent processing modes of different alarm sources and lack of sufficient space-time correlation models, the traditional method often has difficulty in providing accurate fault localization in a large-scale power system. Furthermore, the conventional method cannot flexibly cope with complex factors such as system running state change, equipment aging, load fluctuation and the like, and also cannot extract effective feedback information from historical data to predict and prevent faults. Disclosure of Invention The application provides a secondary equipment fault diagnosis and positioning method and system based on association analysis, which are used for fully capturing complex association between equipment by constructing a dynamic space-time association diagram and combining multi-dimensional information such as time difference, topology weight, alarm level and the like among equipment nodes, so that the limitation of simple dependence on single alarm is avoided. Meanwhile, through multi-window alarm acquisition and iterative computation based on a graph propagation algorithm, the method can effectively reduce errors in fault positioning, and has higher accuracy and reliability in various complex fault modes. Therefore, the innovation point of the application is that a multi-level association and space-time weighting mechanism is introduced, so that the fault diagnosis and positioning are not only based on the current fault information, but also can be adjusted by fully utilizing the historical data feedback, thereby being better suitable for the dynamic operation environment and the complex fault scene of the power system. In a first aspect, the present application provides a secondary device fault diagnosis and positioning method based on association analysis, where the secondary device fault diagnosis and positioning method based on association analysis includes: Step S1, acquiring a topological connection relation of secondary equipment, and calculating a dynamic space-time association weight through a double-exponential decay model according to a topological weight, an alarm time difference, a system running state and a historical feedback factor to construct a dynamic space-time association diagram; Step S2, setting a fast window, a medium speed window and a slow speed window, respectively collecting alarm information, and generating a corresponding fast alarm sequence, a medium speed alarm sequence and a slow alarm sequence; Step S3, mapping the rapid alarm sequence, the medium-speed alarm sequence and the slow alarm sequence to equipment nodes of the dynamic space-time correlation graph, and calculating the correlation strength among the nodes based on the dynamic space-time correlation weight to obtain a rapid correlation sub-graph, a medium-speed correlation sub-graph and a slow correlation sub-graph; S4, carrying out probability iterative computation on the fast associated subgraph, the medium-speed associated subgraph and the slow associated subgraph based on a graph propagation algorithm introducing attention weights to respectively obtain fast fault source probability distribution, medium-speed fault source probability distribution and slow fault source probability distribution; And S5, calculating the window credibility according to the alarm quantity and alarm level of each window, carrying out weighted fusion on the probabilit