CN-122020079-A - Secondary equipment hidden danger intelligent identification method based on wide-area wave recording file

Abstract

A method for intelligently identifying hidden dangers of secondary equipment based on wide area wave recording files includes responding to a fault event of primary equipment of a power grid, carrying out multi-stage station topology analysis based on a power grid topology model, determining surrounding stations related to the fault event, automatically calling wave recording files of the fault stations and the surrounding stations at fault time, analyzing the wave recording files and extracting channel information and waveform image data for hidden dangers identification, and carrying out multi-dimensional intelligent identification on the secondary equipment of the stations through a rule engine and a deep learning model based on the extracted channel information and waveform image data. According to the invention, single-point post analysis is promoted to intelligent screening with wide area linkage and multi-dimension parallelism, and especially AI identification based on waveform images is introduced, so that complex hidden dangers which are difficult to capture by the traditional method can be found, the transition from post analysis to pre-early warning is realized, and the efficiency depth of secondary equipment hidden danger identification is remarkably promoted.

Inventors

• CHEN RUI
• XIN GUANGMING
• CAO TIANZHI
• LIANG HAO
• LI CHANGYU
• LUO JING
• XIA XUE
• YI SHUXIAN
• LIU YINGLIN
• HUANG TIANXIAO
• ZHANG SIQI
• WANG XIAOFEI
• LIU BO
• ZHANG LU
• REN XIANG
• LIU MIAO
• XIE HUAN

Assignees

• 国网冀北电力有限公司电力科学研究院
• 国家电网有限公司

Dates

Publication Date

20260512

Application Date

20251219

Claims (9)

1. 1. The intelligent identification method for the hidden trouble of the secondary equipment based on the wide-area wave recording file is characterized by comprising the following steps of: responding to a primary equipment fault event of a power grid, carrying out multistage station topology analysis based on a power grid topology model, determining peripheral stations associated with the fault event, and forming an associated station group list centering on a fault point; the multi-station wave recording data calling comprises automatically calling wave recording files of fault stations and surrounding stations at fault time based on the associated station group list; Analyzing a wave recording file and extracting channel information and waveform image data for hidden danger identification; based on the extracted channel information and waveform image data, the intelligent identification of the multi-dimensional hidden danger is carried out on secondary equipment of the plant in parallel through a rule engine and a deep learning model respectively.
2. 2. The intelligent identification method for hidden danger of secondary equipment based on wide area wave recording file according to claim 1, which is characterized in that: the hidden trouble identification based on the rule engine comprises sampling loop abnormality identification, action behavior abnormality identification, protection fixed value abnormality identification and opening and closing abnormality identification.
3. 3. The intelligent identification method for hidden trouble of secondary equipment based on wide area wave recording file as set forth in claim 2, wherein the method comprises the following steps: the sampling loop abnormality identification comprises at least one of double sampling inconsistent identification, CT saturation identification and harmonic abnormality identification.
4. 4. The intelligent identification method for hidden trouble of secondary equipment based on wide area wave recording file according to claim 2 or 3, which is characterized in that: The action behavior abnormality identification comprises at least one of trip time abnormality identification and double-set action inconsistent identification.
5. 5. The intelligent identification method for hidden trouble of secondary equipment based on wide area wave recording file according to claim 2 or 3, which is characterized in that: the protection constant value abnormality identification comprises at least one of protection constant value sensitivity coefficient abnormality and minimum action current verification abnormality identification.
6. 6. The intelligent identification method for hidden trouble of secondary equipment based on wide area wave recording file as set forth in claim 5, wherein the method comprises the following steps: the protection fixed value sensitivity analysis comprises the following steps: calculating fault time according to the recording data voltage and current mutation quantity, and calculating an electrical quantity characteristic value at the fault time; Extracting a fixed value from the recorded hdr file, or calling the fixed value of the relay protection device to obtain an impedance fixed value; calculating a protection sensitivity coefficient, namely calculating the ratio of the actual fault quantity to a protection fixed value according to the fault type and the protection principle to obtain a sensitivity coefficient K_sen; Comparing the calculated sensitivity coefficient K_sen with the minimum sensitivity coefficient K_min required by the power system regulation, and performing regulation compliance verification; Analyzing whether the action time sequence and sensitivity of the main protection and the backup protection under the same fault are matched reasonably or not; And outputting the identification conclusion.
7. 7. The intelligent identification method for hidden trouble of secondary equipment based on wide area wave recording file according to claim 2 or 3, which is characterized in that: the on-off abnormality identification comprises at least one of switching value jitter identification and failure loop abnormality identification.
8. 8. The intelligent identification method for hidden danger of secondary equipment based on wide area wave recording file according to claim 1, which is characterized in that: the intelligent identification of the waveform image based on the deep learning model comprises the steps of taking a current waveform image and a voltage waveform image as input, and utilizing a convolutional neural network model to automatically learn and identify hidden dangers, wherein the hidden dangers comprise CT deep saturation characteristic waveforms, PT ferromagnetic resonance waveforms and current waveform distortion caused by breaker operation mechanism faults.
9. 9. The intelligent identification method for hidden trouble of secondary equipment based on wide area wave recording file as set forth in claim 8, wherein the method comprises the following steps: The intelligent recognition of the waveform image based on the deep learning model comprises the following steps: inputting the waveform image into a pre-trained deep learning model, and carrying out feature extraction and pattern recognition on the model; Outputting waveform anomaly type and confidence; and fusing the identification result with the hidden danger identification result based on the rule engine to generate a comprehensive identification report.

Description

Secondary equipment hidden danger intelligent identification method based on wide-area wave recording file Technical Field The invention belongs to the technical field of relay protection and automation of power systems, and particularly relates to a method for realizing intelligent identification of hidden danger of secondary equipment by utilizing fault event triggering, linking multiple stations to analyze recording data and combining with a waveform image artificial intelligent identification technology. Background In a power system, secondary equipment such as relay protection, measurement and control and the like is a key for guaranteeing safe and stable operation of a power grid. When a primary device fails, the associated secondary device should act correctly to isolate the failure. The fault recorder records precious data such as the electric quantity, the switching value and the like of the power grid before and after the fault, and is an important basis for carrying out accident analysis and equipment state evaluation. At present, the investigation of hidden trouble of secondary equipment mainly has the following problems: 1. the post-processing and isolation analysis is usually carried out after the fault occurs, is mostly limited to the recorded wave data of the fault station, and lacks synchronous analysis on the secondary equipment states of the related peripheral stations. 2. Relying on manual experience, the analysis work is highly dependent on the experience of professionals, has low efficiency, and is easy to miss deep hidden dangers and relativity due to fatigue or negligence. In particular, analysis of complex waveform anomalies requires a great deal of time for human interpretation by an expert. 3. The identification dimension is single, traditional analysis is often focused on the direct cause of the fault, and a systematic and multi-dimensional (such as sampling loop, action logic, fixed value matching, open loop and the like) automatic hidden trouble screening means is lacked. 4. The prior automatic method is mostly based on rules and threshold judgment, and lacks effective automatic identification means for complex waveform distortion and transient process abnormality caused by CT (Current Transformer ) saturation, ferromagnetic resonance, breaker mechanical characteristic change and the like. 5. The preventive early warning cannot be formed, namely, due to the lack of a systematic automatic analysis tool, the defects of equipment which do not cause accidents but exist (such as slight abnormality of a sampling loop, slight deviation of protection action time, early mechanical failure characteristics and the like) are difficult to discover in time, and the transition from 'post analysis' to 'pre-warning' cannot be realized. Therefore, a technical scheme capable of automatically, rapidly, multi-dimensionally and systematically identifying the hidden trouble of the secondary equipment across the factory site is urgently needed. Disclosure of Invention In order to solve the defects in the prior art, the invention provides an intelligent identification method for hidden danger of secondary equipment based on a wide-area wave recording file, which aims to solve the problems of low analysis efficiency, single dimension, lack of linkage and preventability in the prior art. The invention adopts the following technical scheme. The invention provides a secondary equipment hidden trouble identification method based on a wave recording file, which comprises the following steps: responding to a primary equipment fault event of a power grid, carrying out multistage station topology analysis based on a power grid topology model, determining peripheral stations associated with the fault event, and forming an associated station group list centering on a fault point; the multi-station wave recording data calling comprises automatically calling wave recording files of fault stations and surrounding stations at fault time based on the associated station group list; Analyzing a wave recording file and extracting channel information and waveform image data for hidden danger identification; based on the extracted channel information and waveform image data, the intelligent identification of the multi-dimensional hidden danger is carried out on secondary equipment of the plant in parallel through a rule engine and a deep learning model respectively. Preferably, hidden danger identification based on a rule engine comprises sampling loop abnormality identification, action behavior abnormality identification, protection fixed value abnormality identification and open-close abnormality identification. Preferably, the sampling loop anomaly identification comprises at least one of double sampling inconsistency identification, CT saturation identification and harmonic anomaly identification. Preferably, the action behavior abnormality recognition comprises at least one of trip time abnormality recognition and dou