A Survey on Intelligent Predictive Maintenance (IPdM) in the Era of Fully Connected Intelligence
Tianwen Zhu, Yongyi Ran, Xin Zhou, Yonggang Wen
Abstract
Predictive Maintenance (PdM) refers to a maintenance paradigm that performs maintenance only after the analytical models predict certain failures or degradations. While traditional PdM has improved maintenance efficiency, its scalability is constrained by network limitations, resource allocation challenges, and the computational demands of AI-driven analytics. With the help of fully connected intelligence, Intelligent Predictive Maintenance (IPdM) overcomes these limitations by leveraging 6G, cloud-edge computing, and AI advancements, enabling a more adaptive and intelligent PdM framework. This survey contributes to the burgeoning IPdM academia from the perspective of communication and networking. Firstly, we explore how communication evolutions and network advancements facilitate IPdM deployment by presenting an integrated cloudedge-device IPdM framework. Under this framework, we analyze the IPdM lifecycle and explore the challenges and solutions within each stage. Particularly, we discuss the novel technical enablers for IPdM in communication an AI. Afterwards, we explore how IPdM helps communication and networking by surveying the applications of IPdM in the regarding fields, such as data centres and sensor networks. Finally, we summarize the crucial future directions for promoting the further applications of IPdM in nextgeneration networking. We believe this work not only contributes to the academic discourse on IPdM but also offers valuable insights for industry professionals navigating the fourth industrial revolution.