Fall risk assessment of cantilever bridge projects using Bayesian network

Abstract

Fall or tumble is one of the most common accidents in bridge construction. Failing to implement safety management and training effectively may result in serious occupational accidents. Current site safety management relies mostly on checklist evaluation; however, its effectiveness is limited by the experience and the abilities of the evaluators, which may not consistently achieve the goal of thorough assessment. Recently, several systematic safety risk assessment approaches, such as Fault Tree Analysis (FTA) and Failure Mode and Effect Criticality Analysis (FMECA), have been used to evaluate safety risks at bridge projects. However, these traditional methods ineffectively address dependencies among safety factors at various levels that fail to provide early warnings to prevent occupational accidents. In order to overcome the limitations of the traditional approach, in this paper a fall risk assessment model for bridge construction projects is developed by establishing a Bayesian network (BN) based on Fault Tree (FT) transformation. The model was found to provide much better site safety management ability by enabling better understanding of the probability of fall risks through the analysis of fall causes and their relationships in a BN. The system has been used to analyze and verify safety practices at five cantilever bridge construction projects currently under construction in Taiwan. It was found that BN analysis is consistent with the conventional safety performance assessment. In practice, based upon the BN analysis by inputting prior information about the site safety management, the probabilities of fall risks and their sensitive factors can be effectively assessed. Proper preventive safety management strategies could then be established to reduce the occurrences of fall accidents at the bridge construction projects.