Experimental Evaluation of Drive-by Health Monitoring on a Short-Span Bridge Using OMA Techniques 123 Fig. 13 Averaged FDD analysis of RAM truck data obtained during DBHM testing at 32.19 kph (20 mph). Demonstrates presence of harmonic frequency intervals and the difference in off and on-bridge spectral content Fig. 14 Short-time analysis of RAM truck DBHM data collected at 32.19 kph (20 mph). Highlights difference between off and on-bridge spectral content using (a) PP and (b) FDD. Bins labeled Harmonic, denote likely harmonics from wheel defects. Bins labeled Vehicle, denote frequencies observed on the vehicle during road tests. Bins labeled Possible Vehicle, denote frequencies either observed on the sprung mass or believed to be detected due to increase in identified peaks during analysis. Bins likely associated with bridge frequencies are labeled Bridge this analysis was decreased to 0.5 Hz. The width of the histogram bins used for displaying cumulative sums was also set equal to 0.5 Hz to ensure that each bin continued to encompass only a single frequency value. As can be seen in Fig. 13, the harmonic frequencies observed in Sect. 6.3 continue to be detected as dominant frequencies by the PP algorithm. To improve the likelihood of detecting lower amplitude bridge frequencies potentially captured in the vehicle response, the number of peak frequencies identified in the averaged and short-time analyses were increased to ten; this is in-lieu of the five peaks previously used in Sect. 6.3. Because bridge frequencies appear to fall within range of the observed harmonic intervals, the previous approach of excluding the interval peaks to identify natural frequencies of the system could not be conducted. Therefore, an alternative approach was taken where frequencies identified in the range of the observed harmonics or vehicle frequencies were labeled as such, while frequency bins that noticeably changed when the vehicle was on the bridge were labeled as suspected bridge frequencies. Figure 14 provides an example of cumulative count data obtained for the front of the vehicle during the short-time PP and FDD analyses, respectively. It should be noted that damping was excluded from these figures to improve readability. A summary of all notable changes in averaged and short-time spectral content for both the front and rear of the vehicle is provided in Table 3. Before the front of the truck enters the bridge in Fig. 14, almost all identified peaks can be labeled as harmonic or vehicle frequencies. When the front of the truck transitions from off the bridge to on the bridge in Fig. 14a, the frequency bin at 15.5 Hz effectively triples. Similar results are observed for the FDD analysis in Fig. 14b, where the 15.5 Hz frequency bin does not appear until the front of the vehicle is on the bridge. A change in spectral content is also observed at 19.5 Hz and 21.5 Hz–22 Hz in Fig. 14b and Table 3. The noticeable change in the 15.5 Hz, 19.5 Hz, and 21.5 Hz–22 Hz frequency bins when the front of the truck is on the bridge suggests they are associated with the identified bridge frequencies in Sects. 6.2
RkJQdWJsaXNoZXIy MTMzNzEzMQ==