Simultaneous Estimation of Wheel-Rail Adhesion and Brake Friction Behaviour

Abstract

In the field of longitudinal train dynamics the brake process is not only a safety critical aspect but it also determines the capacity of the rail network. Therefore, a reduction of the brake distance increases the safety level and the network capacity at the same time. However, to implement an advanced brake control set-up, the knowledge of the wheel-rail adhesion and the brake pad-disc friction is usually necessary. Since the direct measurement of these determining parameters is not reasonable due to technical and economic reasons, the present work presents an estimator framework for their online identification. To ensure a robust and reliable performance of the estimator, a generic wagon model is designed and the observability of the nonlinear system is evaluated. Furthermore, a reasonable synthesis of an extended Kalman filter is discussed that takes account of the system characteristics. In the end, the test results from a roller rig verify the accurate and robust performance of the developed estimator and confirm the great potential of such a concept in the context of mechatronic railway systems.