Review of Machine Learning Based Remaining Useful Life Prediction Methods for Equipment

Abstract

：With the development of science and technology as well as the advancement of production technology, contemporary equipment is increasingly developing towards large-scale, complex, automated and intelligent direction. In order to ensure the safety and reliability of equipment, the remaining useful life (RUL) prediction technology has received widespread attention and been widely used. Traditional statistical data-driven methods are obviously influenced by the choice of models. Machine learning has powerful data processing ability, and does not need exact physical models and prior knowledge of experts. Therefore, machine learning has a broad application prospect in the field of RUL prediction. In view of this, the RUL prediction methods based on machine learning are analyzed and expounded in detail. According to the depth of machine learning model structure, it is divided into shallow machine learning methods and deep learning methods. At the same time, the development branches and research status of each method are sorted out, and the corresponding advantages and disadvantages are summarized. Finally, the future research directions of RUL prediction methods based on machine learning are discussed.

Fault Diagnosis Method for Hydraulic Directional Valves Integrating PCA and XGBoost

“…Due to the availability of big data technology and data mining methods as well as the emergence of new IIoT platforms and machine learning algorithms, fault diagnosis for hydraulic valves based on big data for hydraulic system with condition monitoring is one of the focuses for this research [26][27][28]. Among them, Principal Component Analysis (PCA) is an effective method for dimensionality reduction in big data analysis.…”

State of AI-Based Monitoring in Smart Manufacturing and Introduction to Focused Section

“…Data-driven RUL prediction techniques are typically divided into two groups, i.e., statistical methods [37] and machine learning based methods [38]. Statistical methods are based on the theory of statistics.…”

Remaining useful life prediction via long‐short time memory neural network with novel partial least squares and genetic algorithm

“…They are generally established using multiple heterogeneous variables based on state monitoring parameters during equipment operation. These methods may be “traditional” or deep‐learning‐based according to the learning depth of the structural model 16 . Traditional learning methods, such as neural network‐based and support vector‐based methods, are difficult to automatically process and require the analysis of large amounts of monitoring data due to the limited signal processing ability.…”