Rail corrugation is a significant problem in railway industries worldwide, manifesting as a near periodic wear pattern on the railhead with extended use. This wear pattern causes significant vibrational and sound problems, normally to the point at which periodic regrinding of the railhead is required as a control measure. This is an extremely expensive process and therefore, it is of interest to investigate other alternatives for mitigating rail corrugation on corners. Previous authors have investigated cornering effects on corrugation but not those relating to altering nominal speed or speed distribution characteristics under cornering conditions. Similarly, investigations into speed effects have either been limited to the correlation of train speed to wavelength fixation or variation of speed under straight track conditions.
This thesis presents studies on the effects of altering various statistical train speed measures on the rail corrugation growth process under cornering conditions. Specifically, the primary aims of this thesis were to predict the effects of increasing speed variance on corrugation growth under cornering using time domain modeling, use field measurements to validate time and frequency domain modeling predictions of growth rate and wavelength under variable speed conditions and develop new models to analyze higher order speed effects on the corrugation growth process.
Two dimensional frequency domain and comprehensive three dimensional time domain cornering models were used to predict growth rates of corrugation on a cornered site in suburban Brisbane under investigation. Initial predictions of growth rate under the experimentally measured train speed distribution were validated against intermittent field measurements of the corrugated rail profile. Results showed a 39% reduction in corrugation growth rate from doubling the existing standard deviation of 8.4%. Further validation of frequency modeling predictions of corrugation wavelength was performed at 5 cornered test sites. In the process of collecting field measurements of corrugation profiles, a method of separating two dominant wavelengths of corrugation due to discrete sleeper support was developed and used. Frequency domain modeling predictions of corrugation wavelength, based on experimentally measured receptances and speed distributions, showed very good correlation with field data.
For the purposes of analyzing speed distribution asymmetry effects under cornering conditions, a numerical force balance cornering model was developed and used with the previously used frequency domain model to make predictions of corrugation growth rate under a range of standard skewnesses in speed and nominal speeds for single train passes. Controlling train speed skewness was shown to potentially increase or decrease corrugation growth by up to 19.7% and 12.3% respectively. To isolate the statistical speed effects, a statistical moment expansion corrugation model was developed for straight track conditions and simulations to observe the effects of varying standard skewness and kurtosis were performed. Results showed an increased reliance on higher order effects as the standard deviation was increased and no universal trend in the effects of altering skewness or kurtosis.
The most significant contributions of this thesis are the development of models to investigate standard deviation, skewness and kurtosis effects on rail corrugation growth and prediction of the effects of altering speed characteristics under cornering conditions on dominant wavelength and growth rate using these models. Secondary to this, field validation of modeling predictions of growth rate and dominant wavelength under variable speed over a range of cornered site conditions is provided.