Quantification of uncertainty in the prediction of railway induced ground vibration due to the use of statistical track unevenness data

Environmental vibrations due to railway traffic are predominantly due to dynamic axle loads caused by wheel and track unevenness and impact excitation by rail joints and wheel flats. Because of its irregular character, track unevenness is commonly processed statistically and represented by its power spectral density function or its root mean square (RMS) value in one-third octave bands. This statistical description does not uniquely define the track unevenness at a given site, however, and different track unevenness profiles matching the statistical description will lead to different predictions of dynamic axle loads and resulting ground vibration. This paper presents a methodology that allows quantifying the corresponding variability in ground vibration predictions. The procedure is derived assuming the geometry of the track and soil to be homogeneous along the track. The procedure is verified by means of Monte Carlo simulations and its usefulness for assessing the mismatch between predicted and measured ground vibrations is demonstrated in a case study. The results show that the response in time domain and its narrow band spectrum exhibit significant variability which is reduced when the running RMS value or the one-third octave band spectrum of the response is considered.