# Train vibrations and site effect estimation Using ambient vibration measurement to estimate amplification of vibrations from trains

[Examensarbete på avancerad nivå]

Train induced vibrations are known to cause annoyance and sleeping disorders among people living close to rail roads. This and other problems arising around train vibrations are brought up to date as the current government as well as previous ones suggest building of high speed train routes. The methods, commonly used in Sweden, for predicting levels of train induced vibrations are semi-empirical meaning that a formula is used in which site speciﬁc constants require measurements at the site to determine their value. This thesis aims to study ampliﬁcation of vibrations whose source are trainpassages. Speciﬁcally the fundamental frequency, having a wavelength corresponding to four times the soil depth is subject for the analysis. A method frequently used by seismologists to estimate ampliﬁcation of earth quakes at certain sites is used. By measuring ambient vibrations with a three-axial accelerometer or geophone and dividing the mean of the two spectra of the horisontal components by that of the vertical,the fundamental frequency, f0, will be indicated by a peak in the resulting spectrum, called the H/V ratio. The method can also, to some extent, estimate the ampliﬁcation factor at f0. The effect of f0 on train vibration is studied in three steps. In the ﬁrst step, measurement results from previous studies are studied with regard to f0, derived from soil depth maps and estimated deformation parameters. For the next step, FEM calculations are carried out studying transfer functions at the surface of a soil deposit with soil depth as parameter. Material parameters are chosen so that the fundamental frequency should appear within the studied frequency range. Also,measurements are performed using the H/V-ratio technique. At a meadow adjacent to a rail road, ambient vibrations are measured as well as vibrations from trains passing by close to and further away from the rail in order to obtain a transfer function. The spectra obtained from the train passages at each point, as well as the transfer functions, are compared with the corresponding H/V-ratio to see if there is ampliﬁcation at f0. The results from the literature study indicate that f0 does play a role in that sense that wave propagation below that frequency cannot occur,theoretically. The study of previously made measurements conﬁrms this as sites with a high estimated f0 are not subject to high vibration levels. The FEM calculations also show a dependency of transmitted wave energy on soil depth in that sense that there is a frequency below which there is very weak wave propagation. The response spectra at 10 m and 40 m as well as the transfer function from the source to the ﬁeld point also inherit a peak at this frequency which is one third octave lower than the expected fundamental frequency, concerning the vertical component. As for the measurements, no conclusions can be made as the measurement equipment used was not sensitive enough to capture ambience. Nevertheless, the peaks in the H/V-ratio, although weak, appear at the same frequencies as those at which ampliﬁcation occur in the measurement of train passages. It cannot be concluded, but also not excluded, that the difference in the measured spectra at the different points depend on f0.

**Nyckelord: **train induced vibrations, site effect estimation, ambient vibration, geodynamics

Publikationen registrerades 2019-06-11. Den ändrades senast 2019-06-11

CPL ID: 256724

Detta är en tjänst från **Chalmers bibliotek
**