Bi-Weekly Talk: Matthias Volk and Norman Weik: A DFT Modeling Approach for Infrastructure Reliability Analysis of Railway Station Areas
Wednesday, August 21, 2019, 10:30am
Location: RWTH Aachen University, Department of Computer Science - Ahornstr. 55, building E3, room 9u10
Speaker: Matthias Volk
und Norman Weik
A DFT Modeling Approach for Infrastructure Reliability Analysis of Railway Station Areas
Abstract: Infrastructure failures—in particular in station and junction areas—are one of the most important causes for train delays in railway systems. Individually, subsystems, such as track circuits or radio communication, are well understood and have been analyzed using formal methods. However, verification of the capability of station areas to fulfill operational design specifications as a whole remains widely open. In this talk, we present a fully automatic translation from station area infrastructure to dynamic fault trees (DFT) with special emphasis on field elements including switches, signals and track occupation detection systems. Reliability is assessed in terms of train routability, where feasible train routes consist of the set of train paths projected in the interlocking system including their requirements w.r.t. the state of field elements. Analysing the DFTs by probabilistic model checking techniques allows for new performance metrics based on, e.g., conditional events or the sequence of failures, which can serve to provide additional insights into the criticality of field elements. We demonstrate the feasibility of the DFT-based analysis based on data for railway stations in Germany where the generated DFTs consist of hundreds of elements.