Optimizing condition monitoring of ball bearings: An integrated approach using decision tree and extreme learning machine for effective decision-making

Abstract

Abstract This article presents a study on condition monitoring and predictive maintenance, highlighting the importance of tracking ball bearing condition to estimate their Remaining Useful Life (RUL). The study proposes a methodology that combines three algorithms, namely Variational Mode Decomposition (VMD), Decision Tree (DT), and Extreme Learning Machine (ELM), to extract pertinent features and estimate RUL using vibration signals. To improve the accuracy of the method, the VMD algorithm is used to reduce noise from the original vibration signals. The DT algorithm is then employed to extract relevant features, which are fed into the ELM algorithm to estimate the RUL of the ball bearings. The effectiveness of the proposed approach is evaluated using ball bearing data sets from the PRONOSTIA platform. Overall, the results demonstrate that the suggested methodology successfully tracks the ball bearing condition and estimates RUL using vibration signals. This study provides valuable insights into the development of predictive maintenance systems that can assist decision-makers in planning maintenance activities. Further research could explore the potential of this methodology in other industrial applications and under different operating conditions.