Multi-output Classification for Compound Fault Diagnosis in Motor under Partially Labeled Target Domain

2025-03-15

AI Summary

This paper addresses the dual challenges of label scarcity (i.e., partial labeling) and co-occurring multiple faults in cross-domain fault diagnosis of rotating machinery. To tackle these issues, we propose a multi-output classification-based domain adaptation method tailored for motor composite fault diagnosis. Specifically, we introduce multi-output classification (MOC) into partially labeled domain adaptation for the first time, enabling independent quantification of severity levels for each fault type. We further devise a novel joint alignment strategy integrating multi-kernel maximum mean discrepancy (MK-MMD) and entropy minimization, and incorporate frequency-domain layer normalization (FLN) to better model vibration signal characteristics. Evaluated on six partially labeled cross-domain tasks, our method achieves significantly higher Macro-F1 scores than state-of-the-art baselines, demonstrating superior robustness and generalization capability under both label-scarce and multi-fault conditions.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

This study presents a novel multi-output classification (MOC) framework designed for domain adaptation in fault diagnosis, addressing challenges posed by partially labeled (PL) target domain dataset and coexisting faults in rotating machinery. Unlike conventional multi-class classification (MCC) approaches, the MOC framework independently classifies the severity of each fault, enhancing diagnostic accuracy. By integrating multi-kernel maximum mean discrepancy loss (MKMMD) and entropy minimization loss (EM), the proposed method improves feature transferability between source and target domains, while frequency layer normalization (FLN) effectively handles stationary vibration signals by leveraging mechanical characteristics. Experimental evaluations across six domain adaptation cases, encompassing partially labeled (PL) scenarios, demonstrate the superior performance of the MOC approach over baseline methods in terms of macro F1 score.

Problem

• Addresses fault diagnosis in motors with partially labeled data.
• Improves accuracy by independently classifying fault severity.
• Enhances feature transferability between source and target domains.

Innovation

• Multi-output classification for fault severity
• MKMMD and EM enhance feature transferability
• Frequency layer normalization handles vibration signals