guest ::
TY - THES AU - Liebhold, Alexandra Angelika TI - Evaluation of dispatching algorithms in a railway lab under realistic conditions PB - Rheinisch-Westfälische Technische Hochschule Aachen VL - Dissertation CY - Aachen M1 - RWTH-2024-11302 SP - 1 Online-Ressource : Illustrationen PY - 2024 N1 - Veröffentlicht auf dem Publikationsserver der RWTH Aachen University 2025 N1 - Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2024 AB - As the traffic volume operated via rail constantly increases, delayed trains can severely impact the punctuality of other trains within the network. Train dispatching aims at reducing or avoiding the extent of such secondary delays while increasing the overall punctuality by retiming, reordering or rerouting of trains in the network. Today, dispatching is mostly performed manually by experienced human dispatchers. Over the past years, several decision support systems, which provide recommendations to the human dispatchers, have been developed and successfully implemented during real operation. Fully automated dispatching systems are still rarely applied. Prior to the commissioning of new dispatching systems, extensive testing is required. Theoretical algorithms are first tested for their general feasibility during computer-based case studies and simulations, which are often based on real-world data. After successful simulations, new dispatching systems are normally assessed for their practical usability during field tests on small segments of the real network. However, field tests are extremely costly and long planning horizons are involved. Railway labs can provide an intermediate level testing environment for new dispatching systems under more realistic conditions than pure computer simulations. Tests performed in a railway lab are also more cost-effective than tests performed directly on the real field, while not impeding any real-world operations. However, only few dispatching systems were tested in railway labs so far and previous tests did not incorporate fully automatic operation and thus dispatching decisions were not implemented automatically. To enable automatic real-time train operation under realistic conditions and incorporate dispatching decisions from an external dispatching system during operation, a railway lab control software was extended, enabling automatic train control and train route setting based on a given timetable. Train speed profiles and speed curves were modeled and implemented to allow for realistic driving characteristics and running times. Furthermore, a communication interface to connect an external dispatching system was developed. Via the interface, constant information on the current operational situation are provided to the dispatching system. Dispatching measures are forwarded from the dispatching system to the railway lab control software and implemented during operation accordingly. Predefined disturbance scenarios can be imported into the control software and incorporated during operation. To further allow for energy-efficient train driving, a novel method for the energy optimization of single train runs, based on recommended time corridors for the passing of block signals, was developed and included in the railway lab control software. The developed software systems provide a realistic testing environment for dispatching algorithms and strategies. During a case study, the validity of the developed testing environment was confirmed. The railway lab control software was used to control real-time train operations on the ELVA, the Railway Signalling Lab of RWTH Aachen University, while the external dispatching system OptDis, which is integrated in the software LUKS®, was connected. As disturbance scenarios can easily be reconstructed, the developed testing environment enables the direct comparison of different dispatching strategies. Several disturbance scenarios were constructed and dispatching was performed both automatically by OptDis and manually by an experienced human dispatcher for all test cases.The evaluations show that the developed software systems provide a realistic testing framework for different dispatching strategies, which can serve as an intermediate level testing environment between computer simulations and real field tests. Moreover, the obtained results suggest significant benefits of automated dispatching systems over manual dispatching if operational situations are very complex and decisions must be made quickly. The energy-saving driving method was also tested in a case study on the ELVA network and promising results were obtained. LB - PUB:(DE-HGF)11 DO - DOI:10.18154/RWTH-2024-11302 UR - https://publications.rwth-aachen.de/record/998323 ER -