Fault online prediction method and system based on time sequence feature selection

By using the timing feature selection algorithm in radar fault prediction, the input features of the BP neural network model is solved, and the problem of redundancy and poor prediction performance in the prior art is achieved, and a more efficient fault prediction effect is achieved.

CN119720009AActive Publication Date: 2025-03-28THE 724TH RESEARCH INSTITUTE OF CHINA STATE SHIPBUILDING CORP LTD
4 Cites 0 Cited by

Patent Information

Application Number
CN202411705300.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-26
Publication Date
2025-03-28
Estimated Expiration
2044-11-26

AI Technical Summary

Technical Problem

In the existing radar fault prediction methods, the selection of input timing characteristics cannot be effectively performed, resulting in redundant input features of the model, high spatial complexity, and poor prediction performance.

Method used

A fault online prediction method based on timing feature selection is proposed. By eliminating repeated records in the data set, performing data normalization processing, establishing a backpropagation BP neural network prediction model, and using the timing feature selection algorithm to select the input timing features of the model, reducing redundant features and improving the model prediction accuracy.

Benefits of technology

The timing feature selection algorithm reduces the redundancy of the input feature of the model, reduces the spatial complexity, improves the prediction accuracy and generalization performance of the prediction model, and can more effectively predict the radar fault online.

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

Abstract

The invention discloses an online fault prediction method and system based on time sequence feature selection, and the method comprises the steps: carrying out the preprocessing of original data collected through module monitoring points, and obtaining a state monitoring time sequence data set of M monitoring points; dividing a training set, a verification set and a test set; establishing a fault prediction model for the fault characteristic quantity of each monitoring point of the module, and performing time sequence characteristic selection on the fault prediction model to obtain the optimal input time sequence characteristic of each fault prediction model; the prediction model subjected to time sequence feature selection is trained, and the prediction performance of the model is improved; and performing online prediction on the fault characteristic quantity of each monitoring point through the trained fault prediction model. According to the method, feature selection is carried out on the input time sequence features of the fault prediction model, redundant input features are reduced, the space complexity of the prediction model is reduced, the prediction precision of the prediction model is improved, good generalization performance is achieved, and a basis can be provided for preventive maintenance of radar equipment.
Need to check novelty before this filing date? Find Prior Art