Quantifying Uncertainty in Visual Inspection Data

Abstract

Visual inspection is the most common form of condition monitoring used by bridge owners. Information derived from visual inspection data is commonly used to indicate the performance of bridge stocks and inform bridge management decisions. However, several studies have highlighted that the inherently subjective nature of the methods used to record this data can result in uncertainty, due to differences between different inspectors' perceptions of the severity and extent of defects. It is important for asset managers to understand the nature of this uncertainty and the implications for decision making. This paper reports the results of a study which compared scoring of bridge defects by pairs of independent inspectors across 200 bridge structures on England's strategic road network. A sample of 200 structures was selected to be representative of Highways England's stock with regard to, inter alia, age, condition and structural form. Routine Principal Inspections for these sample structures, undertaken every six years by the relevant maintaining agents, were also attended by inspectors from WSP Ltd, with defects scored independently by each inspector. The results of these comparisons were used to derive an empirical profile of the uncertainty in different individual defect severity and extent scores. Statistical methods were then used to derive empirical probability density functions for the values of bridge and stock level condition metrics according to the widely adopted Bridge Condition Indicator system. The reported results highlight trends in the reliability of individual defect scores and the impact of uncertainty on commonly used performance metrics.

A survey of emerging technologies for the future of routine visual inspection of bridge structures

“…The metrics which result from visual inspections can be used to assess the change of condition at stock or regional level but such processes can be unreliable for detecting the change of condition for an individual structure (cf. Bennetts et al 2018. Undertaking visual inspections can cause network disruption and put inspectors at risk, especially on highways that have been converted to 'all lane running' and therefore always require lane closures even for short duration inspections.…”

OpenIPDM: A Probabilistic Framework for Estimating the Deterioration and Effect of Interventions on Bridges

“…Nonetheless, interpreting information from visual inspection data is known to be challenging [14,4,8]. The challenges arise from technical aspects such as, the subjective nature of the evaluation method [1,3,8]; and non-technical aspects such as, navigating and performing analyses on thousands of structural elements. Moreover, missing data and outliers can impose inconstancies or cause instability in any type of analyses performed using the inspection data [11].…”

OpenIPDM: A probabilistic framework for estimating the deterioration and effect of interventions on bridges

“…Nonetheless, interpreting information from visual inspection data is known to be challenging [9][10][11]. The challenges arise from technical aspects such as, the subjective nature of the evaluation method [11][12][13]; and non-technical aspects such as, navigating and performing analyses on data from thousands of structural elements. Moreover, missing data and outliers can impose inconstancies or cause instability in any type of analyses performed using the inspection data [14].…”