Information Retrieval Framework for Hazard Identification in Construction

Abstract

The repetitive occurrence of similar accidents in construction is one of the most prevalent characteristics of construction accidents. Similar accident cases provide direct information for determining the risks of scheduled activities and for planning safety countermeasures. Moreover, understanding these cases gives laborers the chance to evade and prepare for an expected accident in their workspaces. Researchers have developed many systems in order to retrieve and use past accident cases. Although the developed systems have a clear and limited target (user), most of them were developed under retrieval methods based on ad hoc systems, which can cause inconvenience to users in using the retrieval system. To overcome these limitations, this study proposes an automated information (about past accident case) retrieval system that can automatically search for and provide (as a push system) similar accident cases. The retrieval system extracts building information modeling (BIM) objects and composes a query set by combining BIM objects with a project management information system (PMIS). In this case, a query set is composed based on work, work conditions, and laborers. The suggested information retrieval system compares a query set with past accident cases and retrieves similar cases with their similarity score. The result includes the coordination where similar past accident cases can occur on the current site and the past accident cases are automatically pushed to the highly relevant laborers. Based on the results of this study, the end users (managers and laborers) can excessively reduce query generation. Furthermore, laborers can easily avoid risks by being provided with similar past accident cases that can happen while they work.