Application of Remaining Useful Life Prediction on Railway Traction System to Support Condition Based Maintenance

The railway traction system plays a crucial role in converting electrical energy into mechanical power to move trains, making its reliability and efficiency vital for the safe and smooth operation of railways. Maintaining the traction system is essential for effective railway asset management. Three primary maintenance strategies are used: corrective, preventive, and predictive maintenance. Predictive maintenance (PdM) is considered the most advanced approach, as it focuses on forecasting potential failures by continuously monitoring equipment conditions. PdM integrates diagnostic data, performance data, maintenance history, operating conditions, and equipment design to predict when maintenance is required. A key element in PdM is estimating the Remaining Useful Life (RUL) of components, which indicates how much time or usage remains before failure occurs. Accurate RUL estimation helps operators plan maintenance effectively, preventing unplanned downtime and reducing maintenance costs. The first step in this research involves calculating the Health Indicator (HI), where in this research to get value of HI is determined by multiplying the health condition parameters with a sensor coefficient. By analyzing trends in these parameters over time, the degradation behavior of the traction system can be assessed. The study found that when the HI value drops below 0.8, significant degradation occurs, serving as a reference for initiating maintenance. A value under 0.8 indicates that the system is approaching a critical condition and requires maintenance to prevent failure. By applying data-driven methods for RUL estimation and health monitoring, this study contributes to improving the reliability, safety, and cost-effectiveness of railway operations through optimized maintenance strategies.