Einflussfaktoren und Wirkungsmechanismen zur Lebenszyklusplanung von Straßenbahninfrastruktur [thesis]

Johannes Kehrer, Norbert Ostermann
Operators of urban tram networks are in general mainly funded by public authorities. In order to guarantee an efficient handling of public funds for construction and maintenance of track infrastructure, the entire lifecycle has to be taken into account. Lifecycle costs (LCC) serve as an apporpriate parameter for comparing different strategic approaches regarding the planning of renewal or maintenance measures of tracks.The present work provides a holistic model to determine the optimal
more » ... he optimal maintenance and renewal strategy for a given element of tramway track infrastructure. The database for the model was supplied by Vienna's tram operator Wiener Linien and includes infrastructure, operation and track measurement data. The optimized lifecycles of the individual track elements in any given section are grouped based on operational constraints in order to reach the global LCC- optimum for the respected section, thus enabling a stategic approach to the predictive maintenance for tramway track infrastructure.Relevant input paramters for rail wear and condition degradation taken from topicspecific literature are assessed for their availability of comprehensive data for the given network. In order to program an automatized forecast model, it is essential for the datasets to be network wide and centrally accessible. With a growing database, it should be possible for additional parameters to be integrated into the model and for it to be automatically and continiously calibrated on the basis of the newly obtained measurment data.To meet these requirements, different techniques of machine learning are applied to a preprocessed dataset. The techniques are selected based on their ability to identify and map correlations of the given parameters in order to determine forecast values. Methods selected on the basis of the requirements are examined for their applicability in the given case and optimized in iterative processes. In addition to the prediction of rail wear as the decisive condition parameter, the rail profile shape of the [...]
doi:10.34726/hss.2020.80561 fatcat:pzofxb4wbvgljgx2q7o2hgo4hm