Fault Diagnosis across Heterogeneous Domains via Self-Adaptive Temporal-Spatial Attention and Sample Generation

TL;DR

This work tackles fault diagnosis across heterogeneous domains where health-state category overlap is incomplete, causing distribution shifts that hinder generalization. It introduces TSA-SAN, a two-stage framework that first generates diverse cross-domain fault samples via distribution alignment (DASG) and interpolation-based synthesis (ISS), and then extracts discriminative temporal-spatial features with self-adaptive instance normalization (SAIN) and a temporal-spatial attention mechanism (TSAM). The method demonstrates superior generalization on CSTR and TE benchmarks and a real IPCTF industrial case, outperforming state-of-the-art homogeneous-domain and cross-domain approaches. By enabling robust cross-mode fault diagnosis without exhaustive multi-mode labeling, TSA-SAN offers practical gains for predictive maintenance in complex industrial processes.

Abstract

Deep learning methods have shown promising performance in fault diagnosis for multimode process. Most existing studies assume that the collected health state categories from different operating modes are identical. However, in real industrial scenarios, these categories typically exhibit only partial overlap. The incompleteness of the available data and the large distributional differences between the operating modes pose a significant challenge to existing fault diagnosis methods. To address this problem, a novel fault diagnosis model named self-adaptive temporal-spatial attention network (TSA-SAN) is proposed. First, inter-mode mappings are constructed using healthy category data to generate multimode samples. To enrich the diversity of the fault data, interpolation is performed between healthy and fault samples. Subsequently, the fault diagnosis model is trained using real and generated data. The self-adaptive instance normalization is established to suppress irrelevant information while retaining essential statistical features for diagnosis. In addition, a temporal-spatial attention mechanism is constructed to focus on the key features, thus enhancing the generalization ability of the model. The extensive experiments demonstrate that the proposed model significantly outperforms the state-of-the-art methods. The code will be available on Github at https://github.com/GuangqiangLi/TSA-SAN.

Figures (11)

• Figure 1: Visualization of Notations: example with three domains and four health state categories
• Figure 2: Illustration of homogeneous vs. heterogeneous domain settings in fault diagnosis.
• Figure 3: Construction process of TSA-SAN.
• Figure 4: Structure of TSA.
• Figure 5: Structure of closed-loop CSTR RN237031.