China Safety Science Journal ›› 2017, Vol. 27 ›› Issue (9): 134-139.doi: 10.16265/j.cnki.issn1003-3033.2017.09.023

• Safety Science of Engineering and Technology • Previous Articles     Next Articles

Research on data-driven model for evaluation of construction safety and its application

BAI Xiaoping, PU Tao   

  1. School of Management, Xi'an University of Architecture & Technology, Xi'an Shaanxi 710055, China
  • Received:2017-06-03 Revised:2017-08-18 Online:2017-09-20 Published:2020-11-16

Abstract: In view of a significant element of subjectivity in the construction safety evaluation process, in order to scientifically and rationally evaluate the safety performance of construction the authors established an index system of construction safety based on the method of literature review, field investigation and expert interview. Then, they used the GA and the factor analysis method to optimize the evaluation indexes and establish a data-driven construction safety evaluation model. Furthermore, they proposed a BN-based quantitative evaluation process of this model. Finally, the model was applied to the safety evaluation of a construction project. The results showed that the data-driven construction safety evaluation model can be used to obtain the probabilities of the project at different safety levels and identify the weak links of the construction safety management so that to improve the level of the construction safety management.

Key words: construction safety, evaluation model, genetic algorithm(GA), factor analysis, Bayesian network(BN)

CLC Number: