Structural health monitoring applied to Connecticut's bridge infrastructure.

摘要

Efforts on a variety of bridges throughout Connecticut are underway to determine effective methods of monitoring the health of our nations aging infrastructure. While a large number of analytical methods exist for determining changes in structural behavior of theoretical models and actively monitored bridges, techniques for monitoring bridges through the use of unknown or random excitation has proven difficult. A variety of techniques and methods are presented in this dissertation, in combination with field data from strain gages, accelerometers, tiltmeters, and temperature sensors, to statistically track a bridge's 'health' using existing parameters presented by various researchers, both at the University of Connecticut and other research locations. The introduction of statistics into these existing parameters allows for minimal data manipulation and as close to real-time monitoring as possible. The obstacle of unknown excitations is also overcome through statistics by comparing newly recorded data sets to previously determined Gaussian distributions representative of a collection of responses as opposed to a single value which is sometimes assumed to accurately reflect a bridges response.