Prediction of train-induced vibrations inside buildings using transfer functions

This article presents a study of the use of transfer functions to predict vibrations inside a building due to train-induced ground vibrations. It is proposed that knowing the transfer function of a building, the vibrations inside a similar building can be predicted given a known ground motion outside. A comparison has been conducted between predicted and measured vibrations inside a building in south-western Sweden due to freight and passenger trains. The transfer function was derived by using a stationary vibrating source. The average particle velocity (1 s-r.m.s.) inside the house was calculated with the transfer functions with an average error of −10%, −0.02 mm/s. Prediction was achieved with a standard deviation of 23%, 0.06 mm/s when no filtering of the data was used. Further work is needed to ascertain the accuracy of this method.