Swiss Federal Railways (SBB) is considered one of the most punctual rail operators in Europe. However, this was not enough for the company. A new data analytics software solution now identifies and classifies the reasons for timetable deviations fully automatically, allowing the causes to be systematically and consistently eliminated.
When operational activities leave little time for optimization
Dispatchers at the SBB operational centers respond quickly to disruptions: Through appropriate measures in the Rail Control System (RCS), such as rerouting trains, they mitigate the impacts of disruptions and delays. In these acute situations, the main focus in the operational centers is on the rapid resolution of the disturbance on-site and the efficient reduction of secondary effects on other trains.
Since disruptions often occur simultaneously and in combination, all measures must be immediately recorded in the disruption management tool (ALEA) to inform approaching trains and regional areas of the rail infrastructure as quickly as possible.
Once normal operations resume, the handling of the disruption transitions into a post-processing phase: Dispatchers had to manually assign every schedule deviation of more than 180 seconds to a primary event and record a disruption justification in a third system (ErZu).
This is a manually intensive and error-prone process that leaves little time for dispatchers’ primary task – ensuring smooth train operations. As a result, root cause analysis often falls by the wayside.
Fully Automated Delay Analysis
The “Data Analytics Software Solution” EFA (Erfassung Fahrplanabweichung), implemented since June 2021, represents a significant innovation: It enables the fully automated identification and classification of reasons for schedule deviations based on various data sources, such as unstructured chats, semi-structured XML logs, or structured database contents. By systematically detecting and assigning the causes, which are subsequently addressed wherever possible, dispatchers can now focus entirely on optimizing rail traffic, as the time-consuming analysis of causes is eliminated.
EFA collects terabytes of data daily from various systems and uses Big Data Analytics to fully automate the examination of schedules. Even chat messages from dispatchers in German, Italian, and French are extracted to identify temporal and spatial correlations of delay causes.
To ensure seamless and high-quality complex analysis and processing across different systems, intelligent orchestration of the corresponding Big Data technologies and interfaces is required. The consulting team, led by Gernot Stocker, Detecon consultant and external IT project manager, was responsible for both the development of the analytics strategy and the technical implementation of the project.
Results
EFA provides daily automated, tailored information that initiates data-driven improvements for rail operations and thus benefits Swiss end customers. This leads to significant time savings and increased process efficiency: dispatchers are freed from the need for additional manual documentation and can focus on resolving disruptions, thereby optimizing overall rail traffic.
Automation has significantly increased the validity of the measurements. Now, all schedule deviations are analyzed comprehensively, not just those for which there was previously enough time.
Dispatchers in the operational centers are relieved of up to 15% of their working time through automated disruption qualification, allowing them to concentrate on the continuation of train operations.
Based on the collected data, improvements can be derived. For example, creeping delays are detected and can be addressed in similar situations before they cause issues.
