Physical and numerical modelling of railway induced ground-borne vibration

Ground-borne vibration induced by railway system is a major environmental concern in urban areas. Many researches, usually based on numerical modelling, have been conducted to study this problem. Numerical modelling usually makes many simplifying assumptions to reduce the complexity of the problem. One commonly used assumption is that soil properties are constant with depth. However, various studies in the field of geotechnical engineering have shown that soil properties change with depth due to the effect of increasing confining stress. Ground-borne vibration is hence unrealistically simulated if constant soil properties are assumed throughout a half space or a soil layer. This research focused on studying the soil non-homogeneity effects due to the variation of soil properties with depth on surface and underground railway induced ground-borne vibration. It also explored the use of physical modelling as an alternative of numerical modelling to simulate this problem. The study was performed using physical modelling in a geotechnical centrifuge and by conducting numerical simulations of the tests. The advantage of conducting centrifuge tests on scaled models is that the soil behaviour, which greatly depends on the stress field, is realistically simulated because the in-situ stress field is replicated within the model. Two series of centrifuge tests were conducted to simulate ground-borne vibration from surface and underground railways. To obtain a better understanding of the experiment results, corresponding numerical models were developed to simulate the two series of tests. Boundary effects due to the presence of the centrifuge container on the two series of measurement results were investigated first. The numerical models were then used to examine the soil non-homogeneity effects. Two cases were simulated in each of the numerical models. One assumed that the properties of model soil were uniform. The other case accounted for the variation of dynamic soil properties with depth. The findings from this study suggest that centrifuge modelling is a useful method to study railway induced ground-borne vibration. Reliable and consistent measurement results can be obtained in the centrifuge. This study also addressed an important issue, boundary effects due to presence of the centrifuge container for the tests. The boundary can significantly affect the behaviour of the models, causing fluctuation in the plots of frequency response functions (FRFs) and amplifying model responses. Comparison of the experimental and numerical results demonstrated that the homogeneous models are able to give good estimates of the model behaviour at one location either at the free surface or the interior, depending on the selection of material parameters. However, in order to obtain good estimates of the whole behaviour of the models, it is important to account for the variation of soil properties with depth when studying railway induced ground-borne vibration.EThOS - Electronic Theses Online ServiceGBUnited Kingdo