Characterizing time series of near-miss accidents in metro construction via complex network theory. (October 2017)

• Record Type: Journal Article
• Authors: Zhou, Cheng; Ding, Lieyun; Skibniewski, Miroslaw J.; Luo, Hanbin; Jiang, Shuangnan
• Is Part Of: Safety science. Volume 98(2017)
• Page Start: 145
• Page End: 158
• Publication Date: 2017-10
• Subjects: Near miss -- Complex network -- Visibility graph -- Time series -- Metro construction -- Safety management; Industrial accidents -- Periodicals; Accident Prevention -- Periodicals; Safety -- Periodicals; Travail -- Accidents -- Périodiques; 363.11
• DOI: 10.1016/j.ssci.2017.06.012
• Languages: English
• ISSNs: 0925-7535

Abstract

Highlights: The visibility graph was used to reconstruct the near miss time series into complex networks. The complex networks for near misses have scale-free, small-world features and hierarchical structures. The characteristics of the near-miss time series was uncovered based on complex network theory. The near-miss data from the city of Wuhan, China mainland metro construction was analyzed. Abstract: Although theoretical analysis to near-miss accidents in construction industry has been frequently advocated, the dynamic and temporal characters of time series of near-miss accidents remain unclear, which could not be easily uncovered by analytical representation approaches. To address this gap, the characteristics of the near-miss accident time series and the mechanism underlying the near-miss accidents in metro construction have been explored from the perspective of complex network theory. Mapping time series into a complex network with visibility graph algorithm, temporal characters and dynamics of inter-event time series of near-miss accidents has been revealed through the analysis and discussion of near-miss accident data from the city of Wuhan, China mainland metro construction. All degree distributions of the construct networks, followed by power law, demonstrate that the inter-event time series of near-miss accidents are scale-free. Moreover, the results show they all have small-world features and are highly clustered into hierarchical structures, indicating that the complex phenomenon are generally existed. With the analysis on the near-miss accident time series via complex network theory, practical insights into near-miss accidents and safety management in metro construction are also proposed.