Research on predictive maintenance of equipment based on artificial intelligence

System maintenance is an important means to ensure equipment safety and availability, the condition-based maintenance (CBM) strategy makes maintenance decisions by collecting and evaluating real-time status information of equipment. In recent years, with the rapid development of artificial intelligence (AI), a more novel predictive maintenance (PDM) has gradually become a research hotspot in this field. This article first introduces the relationship between remaining useful life (RUL) prediction and PDM. Then, artificial intelligence based PDM is divided into machine learning (ML) based PDM and deep learning (DL) based PDM. Subsequently, the core features, advantages, disadvantages, and applicable scenarios of ML and DL were introduced separately. Compared to ML, the superiority of DL lie in its automatic feature extraction capability, efficient big data processing capability, and end-to-end learning paradigm, making it perform better in AI based PDM. However, DL has black box problems, poor interpretability, and issues with complex model deployment and poor performance under small sample sizes. Finally, this article discusses the influencing factors of PDM decisions and the challenges faced by PDM applications. Provided scientific and reasonable suggestions for equipment maintenance technology and management personnel.