A model-based proportional-integral observer design for adhesion estimation in the wheel-rail interface: LMI approach

Adhesion estimation in wheel and rail interface is a crucial element as it has a direct effect on the locomotive characteristics and performance. Since the adhesion condition depends on unpredictable factors and dynamic models, its estimation is a complex task. In this paper, a dynamic model for a single wheelset including parameters in detail is developed to improve the accuracy of the model. Then, a model-based Proportional-Integral Observer (PIO) is designed to estimate adhesion during uncertainties and nonlinear parameters. The stability of the proposed PI observer is proved with Lyapunov theory and Linear Matrix Inequality (LMI). Finally, the performance of the method is evaluated via numerical simulation in MATLAB software, and it is validated with a model set-up in the SIMPACK Multibody Simulation (MBS) software.