Estimation Method for Friction Coefficient between Wheel and Rail by a Single-Wheel Creep-Force Model Using Measurement Data of Instrumented Wheelset

Abstract

The purpose of this study is to improve estimation accuracy for motion variables of a wheelset in a railway vehicle calculated by an observer based on a Kalman filter using a single-wheel creep-force model to evaluate a running safety more realistically. To achieve the purpose, this paper proposes a new estimation method for motion variables of wheelset by estimating the friction coefficient between wheel and rail simultaneously. Validity of the proposed method is verified with simulations to test under various running conditions the values of friction coefficient and track curve radius vary. The verification results show that the proposed method is effective to estimate the friction coefficient when track curve radius is smaller than 700m. Moreover, the proposed method manages to solve a problem that estimation results with low accuracy for the motion variables of the wheelset are calculated when the error between a true and an assumed friction coefficient is large.