ARC-GAN: An auditing and reconstructing framework for zero-shot bearing fault diagnosis

Chao Liu

doi:10.1016/j.conengprac.2026.107089

ARC-GAN: An auditing and reconstructing framework for zero-shot bearing fault diagnosis

2026-06-04

0

PRE

AI

C

Chao Liu

袁亮 (Liang Yuan) *

K

Kai Lv *

W

Wendong Xiao

G

Gui Peng

T

Teng Ran

DOI：10.1016/j.conengprac.2026.107089

Original

Original request for help

Save

Abstract

En 中文

• Proposed ARC-GAN, a novel generative framework for zero-shot fault diagnosis. • Introduced a dual-correction mechanism for content and structural consistency. • Designed a Semantic Auditor to align global manifold topology via metric learning. • Developed an Attribute Reconstructor to distill noise-invariant fault features. • Validated robustness against both signal SNR noise and attribute noise.

ARC-GAN: An auditing and reconstructing framework for zero-shot bearing fault diagnosis

Abstract

Journal

Organization

Cited Papers

Citing Papers