科研之友
微信
新浪微博
Facebook
分享链接
摘要
Learning lessons from near-miss reports is a well-known procedure in various high-hazard industries. The construction industry tends to adopt near-miss management systems, but the procedure is relatively new and has not yet been fully explored or understood. Although the management of near-miss reporting systems in construction has been investigated, no effort appears to have been made to suggest a methodology for technically investigating the content of near-miss reported events. This paper reports a study that implemented both qualitative and quantitative analysis methods for a structured investigation of tower-crane-related incident stories (near misses and accidents). The study began by collecting a large number of incident stories (51 accidents and 161 near misses) that were qualitatively analyzed to form an incident database that served as raw material for further quantitative analyses. The database structure definitions contained categories, which in turn contained variables, such as defining the event severity outcomes on a six-point scale from '1' - near miss to '6' - fatality. Further quantitative analysis suggests comparing groups of similar or identical incidents. The groups (clusters) were established by implementing k-means clustering based on the database definitions as variables. The relative risk potential of each cluster was then quantified by comparing each cluster's severity outcome occurrences to those of the other clusters and to the entire database as well. Findings from the analyses suggest that technical failures are the most hazardous risk factors within the tower crane domain.
- 出版日期2017-1
