Savi Rail Performance Model

The Savi Rail Performance Model (SaviRPM) is a sophisticated simulation and performance modelling application developed by Risk Solutions used by the rail industry. It enables performance teams to generate better understanding and management of performance and capacity risk, using cutting edge methods including: big data analysis, agent based modelling, and interactive visualisations.

SaviRPM provides a powerful strategic decision tool helping train operating companies and Network Rail identify where there are potential risks to performance, and remove system fragility from timetables and service operation, creating a more robust and resilient system to improve service performance for passengers and freight customers.

SaviRPM is an agent-based model and has been designed to play out the interactions and potential conflicts between all rail services across a relatively large geographic area, and how they perform when suffering from typical delays. Many simulation runs create a distribution of service performance, from good days to bad days. Exploring these results using Savi’s interactive visualisation tools provides insights into the root causes of poor performance and how these can be tackled. SaviRPM is also used by performance teams to model proposed solutions to improve performance, reporting the likely service performance improvements.

SaviRPM has been used to:

  • De-risk future timetables: by testing the performance of proposed timetable changes and new timetables.
  • Plan the introduction of new services and infrastructure: by adding proposed new services to timetables and testing/preventing any performance impacts.
  • Improve efficiency of timetable changes: through early testing of timetable options to select the ones that perform best for more detailed design.
  • Maintaining service performance with increases in passenger demand: by testing where increases in passenger volumes might require longer station stops to handle boarding and alighting.
  • Finding and dealing with the worst performing services: by looking for trains that cause the most significant delays to other services and passenger arrival lateness.
  • Tackling the incidents that have the highest impact on service performance: by finding the root causes of delays, and cascading reactionary delays, and modelling interventions to find those that deliver the best performance improvement.
  • Improving passenger journey satisfaction: by measuring the number of passengers arriving on time arrivals as well as trains, and modelling ways to decrease passenger journey lateness.
  • Improving service performance with better stock and crew schedules: by modelling stock and crew resources as they are required by each service, observing where delays to inbound services are affecting outbound services, and improving resource scheduling to mitigate this risk.
  • This video is a demonstration of SaviRPM (the Rail Performance Model)

    Summary View

    This visualisation of SaviRPM results provides a way to compare the performance of alternative interventions to improve services.

    Map View

    This interactive map visualisation is used to identify where delays to train services are occurring, so that rail performance teams can design targeted interventions that improve performance.

    Train View

    This interactive visualisation is used by rail performance teams to find the trains that are causing the most delay to passenger journeys, so that they can design targeted interventions to improve performance.

    Timetable View

    This visualisation has been developed in close collaboration with Network Rail to test changes to timetables, and highlight any risks to exceeding network capacity.

    Power BI Visualisation

    SaviRPM data is uploaded securely to the Microsoft Azure Cloud, and is available for further analysis, report and dashboard generation, and self-service business intelligence.

    Sample simple visualisation published to the web using Power BI.