Utilizing a Digital Model-Based Approach for RUL Prediction in an Intelligent Gearbox

This is the authors' abstract. We don't add key points for this paper.

Gears and bearings are critical components in rotating machinery, where their failure can result in substantial operational downtime and increased maintenance costs. This study proposes a novel digital model-based methodology for predicting the Remaining Useful Life (RUL) of gears, offering a demonstrating real-time computational feasibility based on quantified runtime analysis as an alternative to conventional degradation tracking methods that are often time-consuming and experimentally demanding. The proposed framework incorporates high-fidelity gearbox simulations that account for essential design and operational parameters, including gear geometry, bearing configurations, lubrication conditions, housing stiffness, and boundary constraints. Simulations are conducted in compliance with ISO 281 and ISO 6336 standards to compute damage rates on a per-second basis under full-load operational scenarios, enabling the development of a comprehensive damage database while minimizing the need for physical testing. To facilitate real-time health monitoring, embedded sensor technologies compatible with Industry 4.0 are integrated into the gearbox, continuously acquiring data on torque, speed, temperature, and vibration. These sensor signals are processed and visualized through a LabVIEW-based interface, which provides real-time RUL estimation and load spectrum tracking. Validation through endurance testing demonstrates a strong correlation between root mean square (RMS) vibration levels and gear degradation, supporting the effectiveness of the proposed approach. The results confirm the experimentally validated accuracy within the scope of this study, as well as the practicality of the methodology, establishing its potential as a reliable tool for predictive maintenance and structural health monitoring in intelligent gearbox systems.