Bayesian-network-based fall risk evaluation of steel construction projects by fault tree transformation

Abstract

A fall (also referred to as a tumble) is the most common type of accident at steel construction (SC) sites. To reduce the risk of falls, current site safety management relies mainly on checklist evaluations. However, current onsite inspection is conducted under passive supervision, which fails to provide early warning to occupational accidents. To overcome the limitations of the traditional approach, this paper presents the development of a fall risk assessment model for SC projects by establishing a Bayesian network (BN) based on fault tree (FT) transformation. The model can enhance site safety management through an improved understanding of the probability of fall risks obtained from the analysis of the causes of falls and their relationships in the BN. In practice, based on the analysis of fall risks and safety factors, proper preventive safety management strategies can be established to reduce the occurrences of fall accidents at SC sites.