The EU-funded SMaRTE project studied both CBM and the psychological factors of railway usage. SMaRTE is linked to the cross cutting activities of the large scale Shift2Rail (S2R) programme. A high level objective of S2R is to achieve availability and reliability in rail at lower cost – this requires optimal maintenance scheduling. That in part relies on condition-based maintenance (CBM), whereby trains continuously monitor and report their own condition. This permits as-needed maintenance intervention. As S2R also considers increasing the attractiveness of rail, a second aspect is understanding the human psychological factors that deter potential railway users. While the impacts of fares and similar factors have been thoroughly studied, other more subtle factors remain poorly understood. SMaRTE researchers reviewed CBM practices used in other transportation sectors. They investigated techniques for predicting the condition of rolling stock components and systems and developed tools to optimise maintenance decision-making. The team also studied current and future passenger needs. This resulted in a new ‘Smart Journey’ vision, featuring various recommendations for improving user experience.
Project researchers applied statistical and machine learning techniques for analysing condition data. The techniques detect patterns in the data that predict component failure or degradation, providing an early warning. Also investigated were the various techniques contributing to maintenance decision support, including linear fixed models, survival models and a Markov decision process. In practice, the system advises users about the likely effectiveness of various maintenance options. “We tested both sets of tools on rolling stock components,” explains Daniel Johnson, project coordinator. “In the first case, we used the tools on train traction and braking systems, in collaboration with Shift2Rail’s IMPACT-2 project.” This demonstrated the feasibility of the techniques for predicting impending failures, allowing sufficient response time before the failure becomes terminal. A second test study applied various techniques to wheelset condition data. The study demonstrated that the techniques effectively support decisions about complex wheelset re-profiling over the medium term. This demonstrated cost reductions of up to 35 % on wheelset maintenance over its lifetime, or about 3% annually on overall preventive maintenance costs.
“On the human factors side,” adds Johnson, “we delivered a more holistic understanding of the journey and journey planning process. We analysed user satisfaction and the importance of each stage of the journey experience.” Factors included planning/booking/purchasing, journeys to and from stations, experience at stations and experience on board. Researchers concluded that although ticket price is important for passengers, the ability to book journeys in advance and to find a guaranteed seat were more important. So too were security and safety. Passengers expressed greatest dissatisfaction with station car parking and cleanliness, ticket costs, wifi and power connectivity, and the frequency of peak services. Several of the same factors affected recruitment of non-rail travellers to rail. The team assessed the likely impact of all suggested improvements via three case studies, set around Leeds and Manchester in the United Kingdom. Implementation of all recommendations was predicted to improve demand by 25-37 %. Lowering ticket costs alone predicted demand increase by 9-12 %. Yet, ‘soft factors’ such as cleanliness were also expected to be significant. The most important findings of the SMaRTE project will be incorporated into the new Shift2Rail LOCATE project, for application in locomotive bogies. Regarding human experience, the project’s vision will make journeys more pleasant for passengers, while also being profitable.
SMaRTE, passenger, rail, decision support, railway, journey experience, condition-based maintenance, Smart Journey