Extraction method for a wheel rotation average of longitudinal creep force from strain data acquired by instrumented wheelset without the use of rotation angle sensor

This paper proposes a signal processing method to extract the longitudinal creep force from the strain signal that is acquired by an instrumented wheelset in a railway vehicle. The longitudinal creep force is an important physical quantity to discuss the wheel/rail contact condition and the curving performance. Although the instrumented wheelset can detect the longitudinal creep force by the use of the strain signals to measure the wheel load, a wheel rotation sensor such as a rotary encoder is needed to extract the quantitative force data in the conventional signal processing method. This paper proposes a novel signal processing method using the least square technique, which can extract the average value of the longitudinal creep force in a wheel rotation. The numerical tests using dummy signals show that the proposed method has similar frequency characteristics to the ordinary moving average filter in the region of frequency under 10 Hz when the wheel rotation speed is 20 rad/sec. It is also shown that the proposed method can extract the longitudinal creep force from the actual strain data acquired in the running test of real railway vehicle.