Multi-Condition Fault Diagnosis of Dynamic Systems: A Survey, Insights, and Prospects

TL;DR

This paper formalizes multi-condition fault diagnosis (MCFD) and offers a comprehensive review of two main methodological families: single-model-based and multi-model-based approaches. It details handcrafted feature and learned representation strategies, and contrasts fusion-guided versus identification-guided multi-model frameworks, with discussions on evaluation metrics and real-world mechanical and chemical applications. The work highlights current trends toward domain adaptation and generalization, and articulates critical challenges such as class imbalance, interpretability, and unknown operating conditions, proposing avenues like probabilistic modeling and online adaptation for future progress. Overall, the paper provides a rigorous, benchmark-grounded roadmap for advancing robust fault diagnosis under diverse, dynamic operating conditions.

Abstract

With the increasing complexity of industrial production systems, accurate fault diagnosis is essential to ensure safe and efficient system operation. However, due to changes in production demands, dynamic process adjustments, and complex external environmental disturbances, multiple operating conditions frequently arise during production. The multi-condition characteristics pose significant challenges to traditional fault diagnosis methods. In this context, multi-condition fault diagnosis has gradually become a key area of research, attracting extensive attention from both academia and industry. This paper aims to provide a systematic and comprehensive review of existing research in the field. Firstly, the mathematical definition of the problem is presented, followed by an overview of the current research status. Subsequently, the existing literature is reviewed and categorized from the perspectives of single-model and multi-model approaches. In addition, standard evaluation metrics and typical real-world application scenarios are summarized and analyzed. Finally, the key challenges and prospects in the field are thoroughly discussed.

Figures (13)

• Figure 1: Overview of different methods in multi-condition fault diagnosis.
• Figure 2: Schematic diagram of different multi-condition scenarios. (a) Steady multi-condition scenario. (b) Unsteady multi-condition scenario.
• Figure 3: Statistical chart of publication counts for various methods in MCFD based on Web of Science data.
• Figure 4: Proportion of different descriptions for MCFD in the literature.
• Figure 5: Proportion of different frameworks for MCFD methods in the surveyed literature.