Abstract
In North America, common practices in bridge condition assessment include visual inspection and nondestructive evaluation (NDE) techniques, and results are reported as condition ratings of the bridge components. Assigning a specific condition rating to the component is a difficult task, especially when the threshold values defining the borderlines between the different ratings are not specified. These thresholds are subjectively assigned based on the judgment and experience of the inspector or expert, and can influence decisions on maintenance, repair, and replacement (MRR) of bridges and impact their safety and serviceability. Quality inspection data and accurate condition assessment and rating are the basis for determining appropriate MRR decisions. Thus, in this paper, a novel quality function deployment (QFD)-based approach for assessing bridges is proposed to develop an integrated condition rating based on data collected from visual inspection and ground penetrating radar (GPR) technology, while identifying clear thresholds between the different ratings. The k-means clustering technique, used to define the rating thresholds, is one of the unsupervised learning algorithms that solves the subjective determination of the threshold values problem. This work used four case studies on bridges in the Province of Quebec. The integrated condition model produced ratings of 0.48, 0.49, 0.37, and 0.15 for the four case studies. The developed rating model represented by an integrated condition index was validated with an average validity percentage greater than 81%. The proposed method is expected to advance the state of the art for bridge condition assessment and rating by providing an objective means for making proper MRR decisions. (c) 2020 American Society of Civil Engineers.