Research

Share          

Reliability Centered Maintenance Approach to Optimize Train and Track Systems

Rail transport has become an integral part of urban cities with millions of commuters daily. In Singapore, the mass rapid transit system is the backbone of the public transport system. The reliability of rail transport systems requires increasing focus on preventive and predictive maintenance strategies to ensure safe, reliable and customer-centric rail service operations.​
                                     
 
This project at the SMRT-NTU Corporate Laboratory aims to enhance the existing RCM framework via the development of a software, encompassing the reliability centered maintenance (RCM) approach, to assist in the facilitation of systematic improvement of rail reliability (Figure 1). The objectives are achieved via the software for automated data processing, reliability life data analysis, system reliability analysis and FMECA (Failure Modes, Effects and Criticality Analysis). The approach will determine which components are critical, and assist in deciding the type of maintenance strategy to be employed. The life data analysis fits a statistical distribution (e.g. Weibull or exponential) to time-to-failure data, i.e. when a component enters service until it is replaced, to understand when failures tend to occur. In reliability engineering, the lifetime of a population of components is frequently described by a bathtub curve (Figure 2), whose failure rate trends (decreasing, constant or increasing failure rate) can be represented by the Weibull distribution.                
                                                   
Knowing the failure rate trend allows engineers to implement appropriate maintenance strategies via FMECA. For example, components with infant mortality failures should not be replaced periodically; while components showing wearout behaviour (increasing failure rate) need periodic replacement, while considering feasibility and cost.The life data analysis of more components can be put together as a system for simulation (Figure 3). This can reveal overall information on system performance, as well as assessing the criticality of the​ various components. Simulations can consider the effects of different maintenance strategies to improve or optimize reliability.