Engineering model for curve squeal formulated in the time domain
Curve squeal is a type of railway noise that may arise when a railway vehicle negotiates a relatively tight curve. Squeal is common in curves of a radius lower than 200 meters. A single frequency dominates the radiated sound, which makes squeal a very tonal noise. The high number of tight curves in cities and urban areas, the tonal nature and high noise levels, make squeal a serious source of noise pollution. The rising awareness of the impact of noise on public health increases the need to address the squeal problem. Consequently, there is a need for practical squeal simulation tools. The aim of this thesis is to develop a computationally fast squeal model in the time domain suitable for practical use. For this purpose, an existing squeal model is modified. The tangential contact is modelled using a point-contact model, which considers the contact variables in a global manner. This is in contrast to Kalker’s variational contact model which discretizes the contact area into elements. The friction model and contact compliance are defined in a stringent way in relation to Kalker’s model. In this way, the point-contact model is able to describe the transition of contact conditions from full stick to full slip. Although the proposed contact model is steady-state, it performs well at high frequencies. An upper limit of applicability of at least 5 kHz was found in the validation of the contact model within the squeal model. Compared to the classical validation with prescribed motion, the inclusion of the system dynamics puts different demands on the contact model. This indicates that contact models should be validated/compared in conditions that replicate their specific application as closely as possible. The engineering model is completed by implementing an existing model for sound radiation from the railway wheel, the implementation of which is validated against BEM results. A parameter study involving lateral creepage, wheel/rail contact position and friction is performed using the proposed squeal model. The investigated parameters show a strong influence on squeal occurrence and amplitudes. With the wheel being an important factor in squeal, the influence of the wheel modal damping is also investigated. Results indicate that increasing only the damping of the mode excited in squeal may not be sufficient. Squeal may then occur involving another mode with another frequency and amplitude. The amount of modal damping required to prevent squeal is relatively low.
non- Hertzian contact
tangential point-contact model
Room VM, Sven Hultins gata 6, Chalmers University of Technology
Opponent: Dr. Asier Alonso, CEIT, University of Navarra, Spain