Railway wheel squeal as a result of unsteady longitudinal and spin creepage

Abstract

Railway wheel squeal is an unresolved noise challenge facing the railway industry. Of all the noise sources originating from railways, tonal curve squeal is one of the loudest and most disturbing. Wheel squeal results from frictional self-excited vibration occurring in the wheel-rail contact. Solving the problem of squeal requires researchers to work towards an engineering squeal model that can predict squeal comprehensively and in every situation. This will allow for squeal to be resolved during the design stages of new railway systems and railway system components. In South Africa tonal curve squeal originates from the leading outer and trailing inner wheel-rail contacts of Scheffel self-steering bogies underneath empty freight wagons whilst traversing a 1000 m radius curve. Squeal emanating from the trailing inner wheels is especially problematic, occurring at frequencies between 3.8 and 6.5 kHz with very high intensity. Vehicle dynamics simulations have shown that the combination of worn wheel and rail profiles caused by a wheelset tracking error in the bogie cause affected self-steering bogies to over-steer in the test curve. Over-steering causes both wheelsets of the bogie to displace laterally towards the outside of the curve causing high levels of longitudinal creepage. In agreement with on-track measurements, vehicle dynamics simulations showed that the squealing wheels are subject to predominantly longitudinal creepage with little lateral creepage. In contrast, squeal emanating from the leading outer wheel making contact between the flange throat of the wheel and the rail gauge corner occurs at frequencies above 12 kHz reaching into the ultrasonic range. More familiar in the noise environment, the leading outer wheel in some instances also emitted broadband flanging noise. Experimental evidence suggests that the rubbing between the wheel flange face and rail gauge face that causes broadband flanging noise provides positive damping to flange contact squeal, reducing the amplitude of the tonal squeal or eliminating it completely. Mitigation of squeal due to unsteady longitudinal creepage was achieved using a new outer rail profile design. The designed rail profile prevents over-steering of the affected self-steering bogies, eliminating the large longitudinal creepages at the trailing inner wheel necessary for vibrational self-excitation of this wheel. Modelling the longitudinal creep force and spin creep moment as a feedback loop and testing it for stability using the Nyquist criterion showed that both squeal due to unsteady longitudinal creepage and flange contact squeal can be attributed to mode-coupling instability. Crucial to identifying the instability, was modelling the dynamics of a wheel including the effects of wheel rotation subject to moving load excitation; this captures the cross-coupling dynamics between the normal/longitudinal and normal/spin degrees of freedom that is zero for a stationary wheel. In contrast to the current body of knowledge that does not consider longitudinal and spin creepage as sources of instability for squeal, this study shows that both are important sources of instability for squeal. Thus, it is recommended that a future complete model for curve squeal should include longitudinal and spin creepage and the dynamics of a rotating wheel subject to moving load excitation.