Routine procedure for the assessment of rail-induced vibration

Railway induced ground-borne vibration is among the most common and widespread sources of perceptible environmental vibration, adversely impacting on human activity and the operation of sensitive equipment. The rising demand for building new railway lines or upgrading existing lines in order to meet increasing traffic flows has furthered the need for adequate vibration assessment tools during scheme planning and design. In recent years many studies of rail and ground dynamics have produced many vibration prediction techniques which have given rise to a variety of procedures for estimating rail-induced vibration on adjacent buildings. Each method shows potential for application at different levels of complexity and at different stages of a scheme. However, for the majority of the procedures significant challenges arise in obtaining the required input data, which can compromise their routine use in Environmental Impact Assessment (EIA). Moreover, as the majority of prediction procedures do not provide levels of uncertainty (i.e. expected spread of data), little is available on their effectiveness. Additionally, some procedures are restricted in that they require specific modelling approaches or proprietary software. Therefore, from an industrial point of view there is a need for a robust and flexible rail-induced vibration EIA procedure that can be routinely used with a degree of confidence. Based on an existing framework for assessing rail-induced vibration offered by the USA department of transportation (FTA) this project investigates, revises and establishes an empirical procedure capable of predicting rail-induced vibration in nearby buildings that can be routinely applied by the sponsoring company. Special attention is given to the degree of variability inherent to rail-induced vibration prediction, bringing forward the degrees of uncertainty, at all levels (i.e. measuring, analysis and scenario characterisation) that may impact on the procedure performance. The research shows a diminishing confidence when predicting rail-induced absolute vibration levels. It was found that ground-to-transducer coupling method, which is a critical step for acquiring data for characterising the ground, can impact on the results by as much as 10 dB. The ground decay rate, when derived through transfer functions, also showed to vary significantly in accordance to the assessment approach. Here it is shown the extent to which track conditions, which are difficult to account for, can affect predictions; variability in vibration levels of up to 10 dB, at some frequency bands, was found to occur simply due to track issues. The thesis offers general curves that represent modern UK buildings; however, a 15 dB variation should be expected. For urban areas, where the ground structure is significantly heterogeneous, the thesis proposes an empirical modelling technique capable of shortening the FTA procedure, whilst maintain the uncertainty levels within limits. Based on the finding and acknowledging the inherent degree of variability mentioned above, this study proposes a resilient empirical vibration analysis model, where its flexibility is established by balancing the significance of each modelling component with the uncertainty levels likely to arise due to randomness in the system.