Modelling railway overhead line equipment asset management

The Overhead Line Equipment (OLE) is a critical sub-system of the 25kV AC overhead railway electrification system, which is the main method of railway electrification on the British railway network. OLE failures can result in significant delays and pose risks to passenger safety, therefore, inspection and maintenance is undertaken to improve component reliability and uphold the availability of the system. OLE asset management strategies can be evaluated using a life cycle cost analysis that considers degradation processes and maintenance activities of the OLE components. The investment required to deliver the level of performance desired by railway customers and regulators can be based on evidence from the analysis’ results. This thesis presents a methodology for modelling the asset management and calculating the whole life cost of the OLE to allow such analysis to take place. This research has developed a High Level Petri net model to simulate the degradation, failure, inspection and maintenance of the main OLE components in a stochastic manner. The model simulates all the main OLE components concurrently in the same model and fixed time interval inspections and condition-based maintenance regimes are considered. The various dependencies between the different components and processes considered, such as opportunistic inspection and maintenance, are also taken into account. The use of High Level Petri nets allows the processes considered to be modelled in a more accurate and efficient manner in comparison to standard Petri nets. The model is used to calculate various statistics associated with the cost, maintenance requirements and reliability of the individual OLE components and the OLE system over its life cycle. This is demonstrated using an example analysis for a 2-mile section of electrified line, which also describes how the outputs obtained can be used by decision makers to study the performance of the components and the implications of the maintenance strategy evaluated by the model. Finally, a Genetic Algorithm is used in conjunction with the Petri net developed to find the optimum maintenance strategies that result in the lowest total cost of the system. The optimum strategy chosen results in a 15% lower expected total cost and 10% fewer expected failures in comparison to the maintenance strategy currently implemented for the OLE on the British railway network, whilst requiring a similar number of maintenance visits. The methodology presented considers the OLE components and the processes described above in more detail than previous literature associated with asset management and life cycle cost analysis of the OLE. Additionally, the suitability and ways in which Petri nets can be used for modelling the asset management of other large engineering systems, comprised of numerous components with various dependencies, is confirmed. Furthermore, the practical use of the model, as an asset management tool, capable of calculating a comprehensive range of outputs calculated, is demonstrated