Fault detection and diagnosis of air brake system: A systematic review

The air brake system is critical for the proper functioning heavy commercial vehicles, and maintaining its health is crucial to avoid unfortunate accident consequences. Due to the harsh environment and component degradation, safety hazards during operation are inevitable. Therefore, it is necessary to timely detect defects to minimize downtime and maximize safety. Although current maintenance in bus depots often relies on expert knowledge and human resources, intelligent fault detection and diagnosis (FDD) has gained attention in academic and industrial areas. Monitoring air brake systems using various FDD methods has been researched for decades. However, to the best of the author's knowledge, few review articles report on FDD methods for monitoring air brake systems to provide systematic guidance for practical engineers or academic researchers. To address this gap, this work provides a systematic review to understand the current applications and developments of FDD approaches for air brake systems.This paper first presents the operation rules and common failures of air brake systems. Next, it comprehensively reviews and examines existing feature engineering techniques and FDD methods applied to air brake systems. Furthermore, it identifies current issues and potential research directions to attract more in-depth research for intelligent air brake system maintenance. From the overall literature review level, the generic survey architecture with a systematic methodology for air brake systems can be applicable to other safety-critical system domains. From the level of specific method application, the in-depth analyzes of feature engineering techniques and FDD methods applied can also be adopted in other system domains for health condition monitoring.Our work divides the implemented methods into three categories based on the general FDD workflow: model-based, data-driven, and hybrid. Due to the increasing complexity of system structures, the advantages of model-based methods are gradually weakening. With the development of Internet of Things and Artificial Intelligence technologies, data-driven methods are attracting public attention. Hybrid methods have been adopted by combining the advantages of model-based and data-driven methods. This study can provide valuable insights for academic researchers and industry engineers in adopting FDD for air brake system maintenance and other types of system-level maintenance.