"The S12 is running late due to a malfunction" - announcements like this are triggered by system faults in the rolling stock, very often involving door failures. A tiny gap of just a few millimeters causes the monitor to report an open door - and the entire train is immobilized. On account of the tight schedules, such system disorders can easily result in delays nationwide as trains in Switzerland call at 794 stations and stop in hourly or half-hourly intervals - in urban areas even more frequently.
Considering these figures, the SBB is very keen to analyze system faults and to identify the real causes, like operational errors, defective control technology, pneumatics etc., and to prevent future disruption with targeted measures.
The data lab provides us with an overall view of our data - this makes our analytical work more efficient and effective.
Which train systems are particularly prone to outages? Are there routes or stations where faults are particularly common? Which train types are most likely to have door malfunctions? Answers to questions of this nature are very significant for SBB. In the past, the analysis of the causes was extremely complicated. Large amounts of data - for example on failures, delays, operational services, maintenance tasks - had to be summarized manually from various sources, synchronized, linked logically and validated in order to even start with the analysis.
To simplify the process, a data lab was required that would provide quick access to all the information and a user-friendly analysis. Thanks to extensive cross-industry expertise and experience in big data projects, Zuhlke was eminently qualified for the task. The team, consisting of data analysts, data platform specialists and a project manager, developed a data lab and a dashboard.
Zuhlke's data analysts conducted complex data analysis on specific issues. They developed the required data lab in an agile process in close cooperation with the responsible department at SBB. The lab reduced the time for a fault analysis from roughly 5 days to just half an hour. Thanks to an incorporated analysis tool and practical training in its usage by Zuhlke's data analysts, the maintenance specialists can now perform the fault analyses themselves – while devoting all their energies to the content.
The project had a pioneering character for SBB: It could be the beginning of a new decision-making culture based on big data. The new data lab is currently used for diagnostics, but it also provides the basis for future further applications, e.g. for predictive maintenance.
Extensive savings and well-equipped for the future: