One aim of the intelligent train concept is to provide a single maintainer-friendly view of the train. True digital enablement involves the ability to predict failures, highlight hidden problems and issue real-time alerts when failures happen. By using information in a more structured, controlled and repeatable way to plan corrective rail maintenance, engineers can move towards a condition-based maintenance regime, helping to maximise equipment life, increase availability, reliability and reduce costs. The key is to identify potential failure warning signs before a failure occurs, as the repair cost is usually typically far lower to handle pro-actively as opposed to after the failure has occurred.
Train systems generate a huge amount of data – a standard EMU generates in excess of 5,000 signals, equivalent to 2GB of data per train per day, and gathering this data is costly – data transmission could cost thousands of Pounds per train per year. We think that about 0.25% of this data is critical from a rail maintenance perspective, so the secret is smart use of the full data set. This requires the ability to access this data, process it and then have an on-shore back office solution which can visualise and analyse it, develop predictive algorithms and integrate procedures into maintenance management. Data can then be compared to understand how fleets are performing relative to each other and, if presented in a standardised form, allow for direct comparisons across manufacturers. A key part of the savings comes from changes in maintenance practices based on the information being generated from the condition-based maintenance solution.