Discussion. Ambient Vibration Measurements of the Humber Suspension Bridge and Com Parsions With Calculated Characteristics.

Abstract

The team from Bristol University have demonstrated that ambient vibration tests can be used to determine the natural frequencies and mode shapes of the lowest frequency modes of vibration of long span suspension bridges and so check the values used in design.120. Although not mentioned in the Paper, a three-man team from the Building Research Establishment (BRE) also carried out ambient vibration tests on the Humber Bridge at the same time as the team from Bristol University. The BRE approach was slightly different in that signals from thirteen accelerometers were recorded simultaneously for about twelve hours. These thirteen accelerometers were positioned throughout one-half of the bridge to measure vertical motion, and a recording made. The accelerometers were then re-oriented and two further recordings made, one to measure lateral motion and the other to measure torsional motion. The accelerometers were then moved to the other half of the bridge and the process repeated. This enabled the vertical, lateral and torsional motion of the whole bridge to be obtained from just six recordings. However, the results obtained for the natural frequencies and mode shapes of the bridge are in good agreement with those presented in the Paper, particularly for the vertical modes where the results are almost identical." 121. The main reason for using the longest practical recording time was to try to minimize lack of frequency resolution in the analysis: a problem referred to in Appendix 2 of the Paper. The extremely low natural frequencies of long span suspension bridges, coupled with low damping values, can lead to enormous bias errors in the spectral analysis. Indeed, for any of the ambient vibration tests conducted on long span suspension bridges, if the damping values for all the mainspan deck modes are plotted against frequency, a characteristically shaped curve occurs. Figs 14 and 15 show this for the tests by Bristol University on the Humber Bridge and for the tests carried out on the Golden Gate Bridge.4 The lower the natural frequencies of a bridge, the more pronounced is the increase in apparent damping, so the results from shorter span bridges with higher natural frequencies show a similar but less pronounced curve. This phenomenon can be explained as the bias errors in the spectral analysis swamping the true values, in which case, these curves would not represent real damping values. The results of the 1985 tests by BRE also suffer from the same phen~menon,'~ although they

Vehicle Bump Testing Parameters Influencing Modal Identification of Long-Span Segmental Prestressed Concrete Bridges

“…VITs are then conducted using OSP locations to extract modal information for FE model updating. A review of the scientific literature related to the dynamic characterization of bridges under different excitation techniques allows the identification of the most influential parameters [ 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 ], which are bridge typology, main span length (L), excitation source, sensor-to-sensor separation distance, time-series length, and sampling frequency. Figure 5 shows the distribution of the variables as a function of the main span length L. It is possible to observe that the excitation sources most widely used are ambient and vehicles, the sensor separation distance mainly ranges from L/4 to L/9, the sampling frequency ranges from 100 and 256 Hz, and the time-series length ranges from 1.5 and 15 min.…”