A rapid diagnosis method and system for open-circuit faults in charging modules based on an improved migration light gradient lifter.

By improving the transfer light gradient booster model, combining sliding window data segmentation and temporal feature extraction, dynamically adjusting sample weights, and constructing a hybrid loss function, the problems of poor transfer performance and low accuracy of charging module open circuit fault diagnosis under changing operating conditions are solved, and fast and accurate fault diagnosis is achieved.

CN120123872BActive Publication Date: 2026-03-13SUN YAT SEN UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-02-25
Publication Date
2026-03-13

AI Technical Summary

Technical Problem

Existing data-driven open-circuit fault diagnosis methods for charging modules have poor migration performance under changing operating conditions and low accuracy of diagnostic results, making it difficult to meet the needs of real-time diagnosis.

Method used

An improved Light Gradient Boosting Machine (LightGBM) model is adopted. By using sliding window data segmentation and temporal feature extraction, combined with a dynamic update mechanism for sample weights and a hybrid loss function, a transfer fault classification model is constructed to improve the model's generalization ability and diagnostic accuracy under changing operating conditions.

Benefits of technology

It significantly improves the speed and accuracy of open-circuit fault diagnosis of charging modules, meets the needs of real-time diagnosis, and can maintain a high diagnostic accuracy even when the number of samples in the target domain is small.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120123872B_ABST
    Figure CN120123872B_ABST
Patent Text Reader

Abstract

This invention discloses a rapid diagnosis method and system for open-circuit faults in charging modules based on an improved transfer light gradient booster machine. The method includes: first, segmenting the operating data of the charging power module using a moving sliding window and extracting time-domain features to construct a feature vector set; using LightGBM as the base classifier, designing a dynamic update mechanism for sample weights, and constructing a hybrid loss function combining the maximum mean difference and classification error to build a transfer fault classification model; inputting the source domain dataset and a small amount of the target domain dataset into the model for iterative training, dynamically adjusting the sample weights of the source and target domains; and inputting the remaining target domain dataset into the classification model for testing. This invention is applicable to the rapid and accurate diagnosis of open-circuit faults in the power switching transistors of charging power modules under different operating conditions.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Rolling bearing fault diagnosis method based on multi-source weighted ensemble transfer learning

    CN111506862A

  • Rolling bearing fault diagnosis method based on transfer learning

    CN116028876A