An Innovative Robotic Device for the Maintenance on Condition of Railway Pantographs through Structural Dynamics Analysis

Railway pantograph is a suspension mechanism mounted on electrical trains to collect current from the high-voltage overhead-line equipment. Effective verification of such a critical component requires separate tests for inspecting suspension performance and structural integrity. This, in turn, implies that the pantograph is driven by i) a high amplitude low-frequency load, during the suspension test; ii) a fast low-amplitude excitation, to assess structural health. Still, outdoor maintenance sessions suit only the former inspection because portable devices guarantee not-adequate bandwidth. In this Thesis, we present a portable robotic device that has been developed from scratch in partnership with Trenitalia SpA within a research project on Maintenance on Condition. In particular, the robotic device is meant to support the operator during the periodic assessment of the pantographs in outdoor tests, when the train is at rest and disconnected from the high-voltage line equipment. An innovative solution is presented which enables the twofold functionality of the inspection routine. Indeed, a macro-micro actuation combines the large range of force attained by the macro-actuator with the high-frequency capabilities of a micro-actuator while preserving lightweight. A further topic of the present manuscript concerns the adequate characterization of pantograph dynamics. Based on the application requirements, the most suitable vibration index results to be the Frequency Response Function, in the form of single-input-single-output. Nevertheless, nonlinear damping and interaction between actuator and structure cause the response to be sensitive to the actual profile of the force spectrum. Therefore a high-performance force control was contrived within the bandwidth of the excitation. The discussed design and experiments confirmed that accurate excitation transmission is thus achieved. Finally, different kinds of faults were simulated to ascertain the functionality of the device. What is more, the adoption of robotic technology enabled to deepen whether nonlinear behavior enhances the detection capability. To this purpose, a subtle fault-case was simulated, that is the partial loss of preload in a remote bolted joint. The outcome is that even in a restricted low-frequency region, below 5 Hz, slight fault evidence has been recognized with statistical confidence.