| In this thesis, work was completed to validate the accuracy of a developed statistical based damage detection approach. The damage detection approach uses the differences in actual and predicted strains from locations on the bridge caused by heavy five-axle trucks. These differences, called residuals, are then used to construct control charts which compare undamaged and damaged structure data. The validation was achieved by using sacrificial specimens, which modeled damage sensitive locations, were mounted on the bridge and exposed to ambient traffic loads. Different damage levels, simulating cracks or corrosion, were introduced to the sacrificial specimen. Damaged data were compared to undamaged data through the use of control charts and the damage levels were detected. It was also found that damage needed to be close to a sensor in order to be detected by the control chart.;A short sensor attachment study was also completed in this thesis. Issues arose in attaching sensors to concrete therefore multiple sensor attachment techniques were investigated. Through testing in a controlled lab setting it was found that the new techniques were insufficient in transferring strain from the concrete to the sensor. Further study needed to be completed in order to find a satisfactory sensor attachment technique. |
