A hybrid approach for safety assessment in high-risk hydropower-construction-project work systems

The interactive relationships among human factors involved in large-scale hydropower-construction-project management are analyzed and assessed by the data associated with 186 cases of related accident. Many studies have been conducted on human factors influence on construction accidents. The human factor analysis and classification system (HFACS) is developed to establish a rational and applicable index system for investigating human error in accidents. Also, currently the lambda test is used for the correlation analysis of factors; while the factors are assessed by the methods of the decision-making trial and evaluation laboratory (DEMATEL) and the analytic network process (ANP). We extend the HFACS to evaluate the faulty behavioral risk value in this work. The degrees of interaction between independent factors that involved the data used for constructing the direct-relation matrix in DEMATEL are analyzed via the lambda correlation measurement method. The causal graph and calculate results obtained by DEMATEL shows that “safety supervision and inspection” and “organization and responsibility” are the most important factors. Moreover, the empirical study shows that the hybrid method is more suitable and effective than the traditional ANP method. The evaluation model incorporating DEMATEL and ANP takes into account the interaction between factors and their self-feedback, which are more suitable than the traditional method to solve problems with different degrees of effects among clusters. The weights of construction accident factors calculated by the ANP model and the causal graph derived from DEMATEL method both provide suggestions for safety management in the high-risk work systems of several large hydropower-construction projects.