摘要详情

ID / 提交时间

174 / 2025-06-11 19:29:11

标题

Intelligent Fault Diagnosis of Dynamic Equipment in Public Auxiliary Systems Based on Multimodal Deep Q-Networks

关键字

Cross-modal attention mechanism,Multimodal information fusion,Deep Q network,Fault diagnosis; Public auxiliary system

主题及专题

4、System Monitoring, Fault Diagnosis and Prediction

状态

全文待审

作者

Wenhua Zhang / Guangdong Deer Smart Technology Co., Ltd

Zhifeng Liu / Hefei University of Technology

Minggang Tong / Guangdong Deer Smart Technology Co., Ltd

Xin Wang / Guangdong Deer Smart Technology Co., Ltd.

LIAO ZHIQIANG / Guangdong Ocean University

摘要

Abstract—Aiming at the challenges of insufficient single-signal feature representation capability and inefficient multimodal information fusion for fault diagnosis of rotating equipment in public and auxiliary systems, this paper proposes an intelligent fault diagnosis method based on multimodal information fusion depth Q network (MIFDQN). The method extracts multimodal features by designing a cross-modal attention encoder, fusing vibration signals and sound signals, and combining them with an improved stacked self-encoder deep Q-network for fault classification decision. The proposed method uses continuous wavelet transform (CWT) to convert one-dimensional time-domain signals into two-dimensional time-frequency representations; then the cross-modal cross-attention mechanism is used to realize the deep fusion of vibration and sound signals; finally, the fused features are decoded and classified by using the improved stacked self-encoder deep Q-network. The experimental results show that the proposed method outperforms existing methods on the dynamic equipment dataset of public and auxiliary systems.