Ambient Vibration Testing of the Humber Bridge
When opened in 1981 the
Such new technologies bring with them the need re-evaluate the most suitable monitoring solutions and in this instance re-evaluations were necessary on the modal properties of the Humber Bridge alongside the viability of using stand alone recorders with provision timing (via GPS) to provide histories of response within the analysis of such an extended open space structure. The Humber Bridge as designed can tolerate constant motion and bends more than 3 metres in winds of 129 km/hr (80 mph) at which point safety factors emerge but the towers, although both vertical, are not parallel, these being under 50mm further apart at the top than at the bottom.
An international team comprising:
with support from Dr. Ivan Munoz Diaz, Prof. Aleksandar Pavic, Dr. Stana Zivanovic, Mrs. Eunice Lawton, Mrs. Tuan Norhayati Tuan Chik and Mr. Mohammad Muaz Aldimashki from Sheffield, Prof. Alvaro Cunha from FEUP and Mr John Cooper, Mr Peter Hill and Mr Ian Allenby from Humber Bridge Board tested the bridge during the week 14th-18th July 2008 as part of EPSRC funded research project:
EP/F035403/1, Novel Data Mining and Performance Diagnosis Systems for Structural Health Monitoring of Suspension Bridges.
The exercise had several purposes:
- To re-evaluate the modal properties of the bridge and provide a modal model in digital form, that would be used as a baseline for calibration of a finite element model of the bridge
- To evaluate the viability of using standalone recorders with precision timing (via GPS) to provide time histories of response that could be used for operational modal analysis of such an extended open-space structure
- To evaluate the difficulties of operational modal analysis procedures in estimating modal parameters for super-low frequency structures with short data lengths and to evaluate the effect of non-stationary structural (hence modal) parameters on the procedure of gluing mode shape pieces.
The 1985 test used only three Schaevitz LSOC accelero-meters, about 2 km of cable, a two channel spectrum analyzer and a four-channel analog tape recorder. The testing lasted two weeks, and data processing to identify the modes, replaying data tapes through the spectrum analyzer and using the procedure known as ‘peak picking’ lasted about 6 months. The analog tapes can no longer be read and the digital modal description did not survive migration between storage formats in the last 23 years. Hence while the mode frequencies and general form of most of the mode shapes are useful these are not available digitally. Also the resolution was poorer, the damping ratio estimates are known to be heavily biased due to the crude technology, the signal to noise ratios of the sensors are poor and, made worse by the poor dynamic range of the tape recorder, the torsional mode shapes could not be resolved using only three sensors and it was not possible to measure simultaneously points on towers, main span and side spans. So there are several reasons to need an up to date study.
The new study involved new instruments consisting of ten GSR-24’s utilising internal or external accelerometers, all brought together from FEUP and
The testing was divided into 28 measurements spanning five days: For example on day 2, measurements concentrated on the southern part of the bridge, in the direction of the town of
The configuration for measurement setup 9 is shown in Figure 4 the red dots indicate the 10 recorder locations.