Fault trend prediction method for complex power system

By using parallel channel layers and bidirectional long and short-term memory networks in complex dynamic systems to extract fault characteristics and predict future sequence data, the problem of the inability to quickly adjust parameters in the prior art affects the accuracy of fault trend prediction is solved, and more efficient fault trend prediction is achieved.

CN119740003APending Publication Date: 2025-04-01CHINA STATE SHIPBUILDING CORP LTD RESEARCH INSTITUTE 719
0 Cites 0 Cited by

Patent Information

Application Number
CN202411602149.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-11
Publication Date
2025-04-01

AI Technical Summary

Technical Problem

In the prior art, when dealing with sudden failures of complex power systems or changes in equipment operating status, the parameters cannot be adjusted quickly, which affects the accuracy and timeliness of fault trend prediction.

Method used

By obtaining the sequence data to be detected corresponding to the fault characteristic signal in the power system, input them into the parallel first channel layer and second channel layer for feature extraction, the first fault characteristic and the second fault characteristic are obtained. These features are then fed into a continuous two-layer bidirectional long and short-term memory network for sequence feature extraction, predicting future sequence data to determine failure trends.

Benefits of technology

The accuracy of fault trend prediction is improved, timely prediction of future faults is achieved, and the problem of insufficient real-time and adaptability in the prior art is overcome.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a complex power system fault trend prediction method, which comprises the following steps: acquiring to-be-detected sequence data corresponding to fault feature signals in a power system, and respectively inputting the to-be-detected sequence data into a first channel layer and a second channel layer which are parallel to perform feature extraction to obtain a first fault feature and a second fault feature, inputting the first fault feature and the second fault feature into a continuous two-layer bidirectional long-short-term memory network for sequence feature extraction to obtain a hidden sequence feature of the to-be-detected sequence data; and finally, performing future sequence prediction on the hidden sequence features to obtain future prediction sequence data, and determining a fault trend of the power system according to the future prediction sequence data. By means of the fault detection method and device, the defects that in a fault detection method in the prior art, when sudden faults or equipment operation state changes of a complex power system are dealt with, parameters cannot be rapidly adjusted, and prediction accuracy and timeliness are affected are overcome.
Need to check novelty before this filing date? Find Prior Art