Dynamic Properties of Stay Cables on The Penobscot Narrows Bridge

CHAPTER 7. CONCLUSIONS

Vibration testing was performed on the cable-stays of the Penobscot Narrows Bridge, located between Verona Island and Prospect, ME, by manually exciting the cable-stays and measuring the decay with accelerometers and a portable data acquisition system. The vibration data underwent post-processing using various filters. Natural frequencies and damping ratios were extracted for the tested cable-stays. Confidence intervals on the mean were found for the cables in both the first and second modes.

For phase 1 testing, measured first mode frequencies for the fan A cables tested varied from 0.79 to 1.17 Hz, while those for the longer fan C cables varied from 0.71 to 0.81 Hz. Following installation of the external dampers (phase 2), frequencies were similar indicating that the dampers had little effect on frequency. Results for similar fans A and D as well as those for fans B and C compared well indicating that cables with similar unit mass, tension, and details in different locations have the same dynamic behavior. Measured frequencies compared well with theoretical values based upon the string formula and bridge information provided by the designers.

For phase 1 testing, damping ratios for all cables tested varied from 0.10 to 0.39 percent for the first mode and from 0.17 to 0.46 percent for the second mode. As expected, these ranges increased for phase 2 testing after the installation of dampers, from 1.22 to 2.21 percent for the first mode and from 1.41 to 2.48 percent for the second mode. Analysis indicated that measured damping ratios depend on the magnitude of vibration, suggesting the possibility of nonlinear or nonviscous type behavior of the cables.

The effects of aerodynamic damping on total damping were also studied. While a trend between wind speed and total damping is still possible, the total number of test runs for each cable was relatively modest and the wind conditions during the test period were too limited in scope to create a statistically significant regression line for separating aerodynamic from total damping. Future testing that focuses on the extraction of aerodynamic damping should consider more test runs for each cable, preferably during a much broader range of wind speeds and directions than in the testing reported here.

Initial Scruton values, which are used as a criterion to determine effective cable vibration mitigation, ranged from 0.6 to 2.1 before the installation of dampers. This is well below the target value of five for cables with aerodynamic surface treatments, such as the helical fillet used here. On the other hand, final Scruton values following installation of the dampers ranged from 7 to 12 and were well above the target value of 5 for the cables tested.