Assessment of integration method for displacement determination using field accelerometer and geophone data

A conventional French railway track was instrumented with accelerometers and geophones at three depths: sleeper (surface), interlayer (ITL, z=?0.93 m), and transition layer (TL, z=?1.20 m). A linear variable differential transformer (LVDT) was also used to monitor the displacement at the sleeper level. The recorded data allow the integration method (double for accelerometer and simple for geophone) for displacement determination to be assessed. Several questions need to be addressed prior to the selection of an adequate monitoring system: definition of signal filtering processes, influence on results of the different loading wavelengths, repeatability of measurements, train speed and axle load impact and their ranges of validity for each sensor. It was found that the main frequencies that caused more than 95% of the displacement of the monitored materials are in the low frequency range: <25 Hz for trains running up to 200 km/h. For an intercity train, the low frequencies are normally excited by long wavelengths, for instance, those corresponding to the 1/2 coach distance (λ=13.20 m), the bogies distance (λ=6.3 m), and the axle distance (λ=2.8 m). Comparison between the displacements deduced from the records of accelerometer and geophone and obtained from the records of LVDT shows quite consistent results; the mean displacement amplitudes obtained from accelerometers differ by only 20% from the LVDT records. The train speed does not have a strong effect on the obtained differences between sensors. The embedded sensors also gave consistent displacement results for each analysed depth. Moreover, the displacement amplitudes caused by different axle loads (locomotive or passenger coach) are distinguishable for all sensors at all depths. This validates the integration method used for the displacement determination.