Design of Fuzzy Inference System for Condition Monitoring at Wheel-Rail Interface

Abstract

Level of adhesion at wheel-rail interface plays an important role in the stable operation of a railway vehicle. Direct measurement of adhesion is not possible due to the presence of nonlinearities and unpredictable external contaminants. This paper presents an indirect technique to identify the level of adhesion at wheel-rail interface. The proposed identification technique uses the data recorded by inertial sensors mounted on the wheelset of the vehicle. A fuzzy inference system is used to monitor the data recorded by inertial sensors. Changes in the adhesion at wheel-rail interface are reflected in the recorded data. Which are interpreted by fuzzy inference into contact condition information.