The mechanism of railway curve squeal

Abstract

Railway is an energy-efficient means of transport and it is also an important solution to traffic congestion. However, the noise and vibration problems caused by railways attract more and more attention now. One particularly severe problem is curve squeal noise, which is an intense tonal noise that arises in tight curves under certain running and environmental conditions. The mechanism behind this squeal noise is still the subject of controversy. Two causes have been proposed over the last several decades: falling friction and mode coupling. The first one supposes that a decrease of the friction coefficient with increasing lateral sliding velocity occurs and this is equivalent to introducing negative damping into the system, which then feeds energy into the system. In the mode coupling phenomenon, coupling occurs between the vibration in two different directions and energy can be transferred between them.

In this thesis, firstly, these two mechanisms are explored by using an existing curve squeal model to predict the curve squeal in both the frequency domain and the time domain. This model is improved by including a track model based on mass-spring systems, which is more physical and easier to use in the time domain. The results show that both falling friction and mode coupling can lead to instability. Also, the inclusion of the rail dynamics is found to play an important role in the generation of curve squeal. Moreover, it is found that the inclusion of wheel rotation in the model affects the results for different wheels to different extents. To illustrate the findings in terms of wheel mode coupling and wheel-rail coupling instability from this model, several further studies are then performed.

A two-mode model is developed to assess the respective roles of the mode-coupling and falling-friction instability mechanisms in the generation of curve squeal. Different pairs of modes from different wheels are considered. A parametric study is performed including investigation of the effect of the adhesion coefficient, contact angle, lateral offset of the contact point, wheel damping and friction curve slope. Two basic features are found to be characteristic of the wheel vibration in the presence of mode coupling. The first is a frequency shift meaning that the squealing frequency can be different from the natural frequency of the corresponding wheel mode. The second is a phase difference between the vibration in vertical and lateral directions. Previous wheel vibration measurements are used to give a qualitative comparison with the model to identify these features.

To study the effect of rail dynamics on curve squeal, the rail is firstly modelled as an infinite beam over a continuous elastic foundation. In contrast with the wheel, it is not characterised by vibration modes. Results show that the presence of the rail plays a role in the instability. Various effects are then considered that may change the resonant behaviour contained in the rail dynamics. These include the effect of varying the rail pad stiffness, the influence of the periodic support of the rail, reflections between multiple wheels on the rail, rail cross-section deformation and the inclusion of rail cross mobility. Finally, a reduced model is developed to identify the essential elements of the dynamic behaviour of the rail that can cause instability. In this model, a single wheel mode is included and the rail is represented as a mass, a spring or a damper. It is found that it is not necessarily the introduction of ‘modes’ in the rail that causes the wheel modes to couple with the rail; instead the equivalent mass and/or damper behaviour of an infinite rail is the origin of a wheel-rail coupling phenomenon.

Finally, a laboratory measurement is performed by modifying an available machine originally designed to perform pin-on-disc friction measurements. By using a 1:5 scale model of a railway wheel, squeal noise is observed at two different frequencies. During the measurements the wheel is stationary and is set in the vertical plane while the rotating disc lies in the horizontal one. The axis of the wheel is tangent to the rotating disc. Lateral force and wheel vibration in radial and axial direction are recorded. From the vibration data it is found that the response of the wheel in the vertical and lateral directions are almost in phase and that the squealing frequencies is always almost coincident with a natural frequency of the wheel. For the friction, a mild falling trend can be observed when the sliding velocity increases. For sliding velocities below 0.15 m/s the peak axial vibration velocity is found to be equivalent to the velocity of the rotating disc at the contact point. These observations suggest that stick-slip and/or falling friction can be responsible for the squealing in this testrig while there is no evidence of mode coupling in this specific situation.

More information

Identifiers

ORCID for Giacomo Squicciarini: orcid.org/0000-0003-2437-6398

Catalogue record

Export record