Abstract
It is one of the important issues to investigate the vibration environment of railway bogies, since the vibration of the bogies may result in loosening bolts which fix the parts to the bogie frames or/and fatigue fracture of the parts themselves. A simple technique for predicting the vibration of bogie parts has been proposed in which the acceleration power spectral densities (PSDs) at evaluated points are predicted with the identified frequency response functions (FRFs) between the axle boxes and the evaluated points, together with the measured accelerations of axle boxes. Stationary excitation tests were conducted in the previous studies to identify the FRFs, and the axle boxes or rails were hit with impulse hammers to excite the bogies, while the new approach without the stationary tests is proposed in this study. The FRFs are identified with the accelerations acquired in the preliminary running tests in car depots. The proposed technique are applied to the vibration prediction of the bogies for two types of railway vehicles: an electric car and a diesel car, and the differences or ratio between the predicted and actually measured PSDs are evaluated. It is confirmed that the prediction accuracy can be improved in the case where not only the vertical but lateral and longitudinal accelerations of axle boxes are considered as the excitation inputs.
