Label-Free Cross-Load Soft Fault Diagnosis for DC-DC Converters

View at publisher Summarise with VIRAT
Authors: Guoqing Zhang, Yang Liu, Zhiliang Chen, Yanchen Tu, Yi Zhang, Haoran Wang
Cited by: 0
References: 17

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

DC-DC converters are widely used in power electronic systems, but long-term operation will lead to device aging and parameter drift, which will lead to soft faults and threaten the safety and reliability of the system. Under the conditions of low sampling rate, limited computing power and small samples, it is a major challenge to realize robust and cross condition generalization soft fault diagnosis. To solve this problem, this paper proposes a soft fault diagnosis method based on extreme learning machine (ELM), which improves the generalization ability of the model under different working conditions through the mechanism of pseudo label generation and maximum mean difference (MMD) feature alignment. This method does not need complex network structure, training is efficient, and is suitable for embedded environment deployment. Simulation results show that this method still has strong diagnosis performance and robustness under the condition of small samples and low sampling rate, which provides a new idea of low cost and rapid implementation for soft fault diagnosis of DC-DC converter. Index Terms—DC-DC Converter, Fault Diagnosis, Extreme Learning Machine, Transfer Learning, Power Electronics.

Fields of Study & Subjects

Fault Detection and Control Systems Advanced DC-DC Converters Machine Learning and ELM

Publication details

Journal / source
IEEE PES International Meeting (PES IM)
Pages
1–5
Publication date
18 January 2026
Language
EN