Non-smooth modelling of drive wheels stick-slip

Abstract Motor railway vehicles necessitate enhanced control of wheel-rail contact mechanics to ensure optimal adhesion. During train running, driving axles can exhibit torsional vibrations that compromise adhesion and potentially lead to axle damage. Consequently, the development of dynamic models for analyzing driving axle stick-slip phenomena and control strategies is an area of significant research interest. This work focuses on demonstrating the efficacy of non-smooth models in studying the torsional stick-slip behaviour of axles in traction vehicles during acceleration. Despite their application in various railway vehicle problems, non-smooth models haven't been explored as an alternative for analyzing stick-slip. The proposed model prioritizes simplicity while capturing the essential friction characteristics, enabling efficient dynamic simulations. A set-valued friction law is incorporated, and the equations of motion are formulated as a switch model. Numerical integration is achieved through an event-driven algorithm. The paper showcases application examples for the model. A direct comparison with a “classic” model yields that the non-smooth integration is 5 times more efficient in the stick phase.