PSI - Issue 17

Paulo Morais et al. / Procedia Structural Integrity 17 (2019) 419–426

422

Paulo Morais et al. / Structural Integrity Procedia 00 (2019) 000 – 000

4

expected at that location (Fig. 2). For these tests, the instrumented vehicle had an attached trailer, related with its original intended job.

a)

b)

Fig. 2. Site monitored at PK 63+472: a) details of the piezoelectric accelerometers placed in railway line; b) passage of instrumented vehicle.

Analyzing the results obtained with the on-board instrumentation, over the monitored railway section, it was possible to verify that: • The data produced by the two central displacement transducers installed on the chassis, for normal vehicle operation speeds, does not have adequate quality to directly assess railway stiffness. This problem is related with the location of the transducers on a chassis part that exhibited an unpredicted mobility due to the existence of an additional suspension system between chassis and the engine. This problem can be easily circumvented in the future by relocating the attachment points to other parts of the structure and resume the initial concept of directly measuring railway vertical deformation and stiffness; • The suspension displacement transducers installed still allowed the detection of a wide range of events, permanent and transient in nature, on the vehicle itself or on the railway, thus constituting a vital source of information to better understand the data obtained with the remaining instrumentation (Fig. 3). Examples are the presence of a small wheel defect on the rear shaft and the impact of a sudden stiffness transition from USP type II to USP type I at 57 m mark, on the selected railway section; • The accelerometers installed in the wheel boxes provide mainly qualitative information on the wheel-rail interactions, which coupled with geo-referenced data provided by GPS, allows to pinpoint out the presence of notable events on the railway line, including defects on the railheads or fractured sleepers; • The accelerometers installed in the cabins provided detailed information, with adequate quality, to identify the dynamic features of the various elements of interest, regarding both the vehicle and the railway line (Fig. 4). Again, it can be observed the big impact of the sudden transition of railway vertical stiffness at 57 m mark on both shafts, which are 7.6 m apart. Fig. 5 presents the spectral amplitudes of a Fourier transform applied on the vertical channel of each cabin accelerometer, over four periods of consecutive data acquisition, obtained in the previously described instrumented railway section. Each period corresponds to parts of the railway line with different known dynamic characteristics, respectively: a) bridge; b) bridge equipped with USP; c) railway line over embankment equipped with USP and d) railway line over embankment with no USP. These graphics identify the most likely sources for the observed response peaks, which can be separated into two groups. The first group includes the frequency peaks that depend on the speed of the vehicle, namely: Wheel Per. – frequency associated with the presence of a verified defect in one of the vehicle’s wheels (wheel perimeter 2.636 m); Sleepers – frequency associated with the passage over the sleepers (0.675 m apart); Engine – Engine rotation frequency and harmonics. The second group includes the