An outline of Prognostics and health management Large Model: Concepts, Paradigms, and challenges

Prognosis and Health Management (PHM), critical for preventing unexpected failures and ensuring task completion of complex systems, is widely adopted in the fields of aviation, aerospace, manufacturing, rail transportation, energy, etc. However, PHM’s developments and applications have been seriously constrained by bottlenecks like generalization, interpretation and verification abilities. Large Model (LM), a typical and powerful representation of generative artificial intelligence (AI), heralds a technological revolution with the potential to fundamentally reshape traditional technological fields. Its strong generalization and reasoning capabilities present opportunities to address those PHM’s bottlenecks existing. To this end, by systematically analyzing the current challenges and bottlenecks in PHM, as well as the advantages of Large Model, we propose a novel concept and corresponding three typical paradigms of PHM Large Model (PHM-LM) by the combination of the Large Model with PHM. Additionally, couples of feasible technical approaches for PHM-LM within the framework of the three paradigms are provided to address core issues confronting PHM and to bolster PHM’s core capabilities. Moreover, a series of technical challenges throughout the entire construction and application process of PHM-LM have been deeply discussed for further research recommendation. The comprehensive effort herein offers a comprehensive PHM-LM technical framework, and provides avenues for new methodologies, new technologies, new tools, new platforms and applications of PHM, which also potentially innovates design mode, research & development mode, verification and application mode of PHM, i.e., from traditional customization to generalization, from discriminative approaches to generative methods, and from idealized conditions to practical applications.