Abstract

Bridge condition assessment is one of the most essential elements of Bridge Management Systems (BMS). This is owing to the fact that available inputs from assessment reports are constantly interpreted for maintenance decisions and budget allocation to critical bridges within a region’s inventory. Thus, performing effective bridge assessment is vital to ensure safety and sustainability of the bridge infrastructure. In practice, the evaluation of concrete bridges is mostly conducted on the basis of visual inspection, which is associated with considerable subjectivity and uncertainty inherent in human judgments. Additionally, current bridge assessment practices were found to be oversimplified, with conclusions being often drawn in absence of in-depth review and consideration of critical factors.
To remediate the existing shortcomings and ameliorate the bridge assessment process, this study proposes a fuzzy Hierarchical Evidential Reasoning (HER) approach for detailed condition assessment of concrete bridges under uncertainty. The essence of the suggested framework addresses the treatment and aggregation of uncertain measurements of detected bridge defects, in a systematic manner, to establish an enhanced platform for reliable and detailed bridge assessment. The significant features of this methodology can be summarized in the following points. First, the proposed approach utilizes a generic hierarchy that models the several levels of a concrete bridge under assessment; namely: bridge components, elements, and measured defects. Second, the proposed model is set to account for relative importance weights of all assessment factors in the hierarchical breakdown. Third, a novel HER assessment belief structure is employed to grip probabilistic uncertainty (ignorance) in bridge evaluation, whereas fuzzy uncertainty (subjectivity) is processed through a set of collectively exhaustive fuzzy linguistic variables. Forth, Dempster-Shafer (D-S) theory is eventually applied under the proposed HER framework for the purpose of accumulating supporting pieces of evidence in a comprehensive manner. The suggested model is implemented to arrive at detailed and informative bridge element condition ratings through data acquired from two case study bridges in Canada. As it benefits from a data oriented and structured algorithm, the developed defect-based model is believed to introduce a great deal of objectivity in an otherwise subjective area of infrastructure assessment. This falls within the ultimate goal of enhancing overall public safety and well-being.