Rail joints have been in existence for a long time of which their design has remained primarily unchanged over this period. Rail joints are a typical example of bolted joints. Like other bolted joints, their integrity depends on the quantitative representation of the contact pressure distribution at the interface during design. In service, rail joints are subjected to complex operating stresses, and they demand high maintenance cost because they are safety critical and have the lowest service life of the components on the rail tracks. In this study, non-intrusive ultrasonic techniques have been employed to investigate the parameter relevant to their design, operation and condition monitoring.
The effect of variation in plate thickness and diameter of the bearing surface of the bolt head on the contact pressure distribution at bolted interfaces under varying axial loads was investigated. While it was observed that the contact pressure at the interface increases as the applied load increases, the distance from the edge of the bolt hole at which the distribution becomes stable is independent of the applied load on the bolted joint. However, the contact pressure distribution was observed to vary with the plate thickness. Although the variation in the peak value of average contact pressure distribution in bolted joints does not depend on the plate thickness, the distance from the edge of bolt hole at which the value of the distribution becomes stable increases as the plate thickness is increased. It was revealed that the peak value of the contact pressure distribution decreases as bearing diameter of the bolt head increases, and that the distance at which the normalised average contact pressure distributions become fairly constant also increases as the bearing diameter of the bolt head increases. In the majority of the cases, the distance falls between 3 and 4 of the bolt radius from the edge of the bolt hole. It was also observed that the edge of the bolt head has a pronounced effect on the position of the peak value of the contact pressure distribution at the interface. Furthermore, a model based on a Weibull distribution has been proposed to fit the experimental data, and a good correlation was observed.
Non-intrusive experimental techniques were simultaneously used to investigate the relaxation of contact pressure and loosening of bolted joints subjected to cyclic shear loading. Three critical areas: the contact interface of bolted component, the bolt length and the rotation of the bolt head, were monitored during loosening of the joints. The results show that loosening of bolted joints can be grouped into four stages. The early stage of the loosening of bolted joints is characterised by cyclic strain ratcheting- loosening of the bolted joint during vibration without rotation of the bolt head. The higher the rate of relaxation at this early stage the lower is the resistance of the bolted joint to vibration induced loosening of bolted joints. Furthermore, the rate of loosening at the bolted joint interface is not the same but increases away from the bolt hole. While the rate of loosening of bolted joints largely depends on the amplitude and the number of cycles of the applied dynamic shear load, it is independent of the frequency of the applied load. In addition, increasing the bolt torque was found to increase the loosening resistance of the bolted joint. When joints are subjected to a constant shear load in addition to the dynamic shear load, the loosening rate increase, and this rate depend on the magnitude of this constant shear load.
A normal incidence pulse-echo ultrasonic technique was used to monitor de-bonding at the interface of adhesive bonded insulated lap joints and insulated block joints, subjected to a shear load induced failure. The results revealed that the insulated joint exhibited elastic behaviour before a sudden failure (rupture) of the joint. The de-bonding of adhesive/insulating layer on the web sides of IBJs was found to occur earlier than at any other parts of the joint when the applied load was only a fraction of the peak of the shear load. However, the de-bonding at the top and foot of the rail occurred almost at the peak of the shear load. This same technique was used to monitor the degradation, and eventual failure of IBJs subjected to cyclic shear loading. The results showed that the degradation of the adhesive insulating layer has commenced, and was in progress when the joint was virtually intact and displayed elastic behaviours. The failure at the adhesive interface is indicated by a sudden change in the value of the measured reflection coefficient. After this, the failure of the joint is preceded by plastic behaviour of the joint. In addition, the change in the length of the bolts can also be monitored directly, as a complete failure at the interface of the bolted joints can be highlighted before the failure occurred. The results of this study have shown that, with further development, ultrasound can be used to monitor the condition of IBJs while in service
