Partial sensing information-driven threshold cyclic update graph autoencoder for mechanical anomaly detection

Yi Gao

doi:10.1016/j.ress.2025.111558

Partial sensing information-driven threshold cyclic update graph autoencoder for mechanical anomaly detection

2025-08-06

0

PRE

AI

DOI：10.1016/j.ress.2025.111558

Original

Original request for help

Save

Abstract

En 中文

• Propose TCUGAE as an unsupervised method for multi-condition anomaly detection. • Design the PSG to select high-quality sensor data for graph construction. • Construct the PSGAE model to detect anomalies via reconstruction composite loss. • Develop a TCU strategy for adaptive threshold setting across multi-condition.

Keywords:

TCUGAE

PSG

PSGAE

anomaly detection

adaptive threshold

Partial sensing information-driven threshold cyclic update graph autoencoder for mechanical anomaly detection

Abstract

Journal

Organization