Abstract

The discontinuities in surface profiles of railway wheels, commonly known as wheel flats, are known to impose excessive impact loads at the wheel-rail interface. Such impact loads can cause premature fatigue and failure of the vehicle-track system components, and impede the operational safety. The safe and cost-effective operations of railways thus necessitate continuous monitoring and control of impact loads induced by wheel defects. In this dissertation research, an adaptive wheel-rail contact model is developed to predict contact geometry and impact force as a function of flat geometry, speed and normal load. The model employs radial contact springs and could simulate for either single or multiple wheel flats. Unlike the commonly used Hertizan nonlinear models, adaptive model predicts the contact geometry involving either total or partial contact in the presence of a wheel defect in the contact patch. The proposed contact model is integrated to a roll plane model of vehicle and a three-dimensional flexible track model to derive a coupled vehicle-track system model. The vehicle is modeled as a six-DOF lumped mass system including carbody, bolster, sideframe, wheelset, and primary and secondary suspensions. The track system model considers two parallel Timoshenko beams periodically supported by lumped masses representing sleepers. The rail-pad and ballast are also included through linear visco-elastic elements. Central finite difference technique is employed to solve for the coupled partial and ordinary differential equations of motion for the continuous and discrete system models, respectively. The dynamic response of the wheel-track system is initially investigated under a constant moving load to examine validity of the model and the numerical method. The impact force response of the adaptive contact model in the presence of a single wheel flat revealed reasonably good agreements with available measured data. This agreement was better than that provided by the well-known Hertizan nonlinear point contact model. The results further revealed that discrete sleeper supports act as sources of excitations. The results attained from the parametric study revealed that the normal load, speed and flat size are the primary factors that affect magnitudes of impact forces, while the suspension parameters show only minor effects. Some of the parameters of the track system also revealed important effects on magnitudes of impact force. The coupled vehicle-track system is further analyzed to derive the impact force properties for different wheel flats, operating speeds and loads. The analyses were also performed for single as well as two flats within the same or two opposite wheels of a wheelset. The results suggested that magnitudes of impact forces attributed to the second flat were strongly affected by responses to the preceding flat. The resulting peak impact force may be either higher or lower than that caused by a single flat, depending upon flats geometry, relative coordinates of the flats and operating speed. The results further suggest that the length of a flat, which is regarded as the removal criteria by AAR and Transport Canada, is not sufficient for cases involving either single or multiple flats