15 Ambient Vibration Test of Granville Street Bridge Before Bearing Replacement 155 Fig. 15.3 Expansion joint at the Pier 7 free to display only in longitudinal direction. Support conditions between the superstructure of the bridge and each pier: (a) Pier 1: free to move in longitudinal direction and rotation about vertical direction, (b) Piers 2, 3 and 6: fixed in all directions, (c) Pier 4: left support is free to move in longitudinal direction and rotation about vertical, and the right support is fixed in all directions, (d) Pier 5: both left and right supports are fixed in all directions, (e) Pier 8: free to move in longitudinal direction and rotation about vertical direction 15.3 Description of the Ambient Vibration Test The first phase of the AVT of the Granville Street Bridge carried out on Thursday June 12th, 2013. The test started 10:10 in the morning and lasted at 6:18 in the afternoon of the same day. The site temperature during the AVT was recorded between 13 to18ıC: 13ıC at 10:00 am, reached to 18ıC at 03:00 pm, and down to16ıC at 06:00 pm. Two concrete approaching viaducts, one at each end of the bridge, are included in the AVT. The length of south approach is 64 m and the north approach is 44.6 m. One of the nine sensors is used as a reference sensor located in the main span of the bridge. The remaining eight sensors are divided into two groups of four. Each of the groups is used exclusively on one side of the bridge: one group on the west side and the other on the east side. Two groups of sensors and the reference sensor located in the main span forms one setup of the AVT. Starting from the south end of the bridge, setups progress to the north end of the bridge. Using 13 setups of this kind, the entire bridge is monitored by 106 measurement points. The measurement points within each expansion joint are placed in equal distance. In each setup the north direction of each sensor is oriented to the longitudinal direction of the bridge (Fig. 15.4). The test duration for each setup is 30 min, and the sampling rate of the recordings is 128 Hz. Low gain velocity measurement components of the sensor are used for both data analysis and modal identification. An external GPS attached to each sensor allowed the synchronization of sensor recordings both within each setup and between the setups. The reference sensor is set up to run in continuous recording mode during the entire AVT, but it malfunctioned, and as a result of this some of the recorded data on the reference sensor had been overwritten. Only the last 3 h of data were able to be retrieved from the reference sensor: from 16:04 to 18:18 pm. Such a short duration of data made it impossible to calculate the mode shapes of the entire bridge. In addition to that, due to the technical difficulties beyond our control at site, no measurement had been conducted at measurement point 103, which is located at the concrete viaduct at the north end of the bridge. 15.4 Data Analysis The noise components of the ambient vibration data, by its nature, usually appear as a random phenomenon in the data, whereas the response of the bridge is not random, but consistent at certain frequencies due to the resonance effect of the bridge. One, therefore, can minimize the effect of noise components by using statistical tools such as segmentation (or windowing), averaging and smoothing. To do this, the data is divided into equal lengths of segments (running windows) as indicated in Fig. 15.5, and each running window is shifted from the beginning of the data to the end by a predefined overlap
RkJQdWJsaXNoZXIy MTMzNzEzMQ==