Factors Analysis And Assessment In High-risk Operation System

With the rapid development of China’s economy, more and more large-scaleconstruction projects constantly appeared. Work safety problems have becomeincreasingly apparent.Construction accidents occurred in high-risk jobs broughtincalculable loss to construction workers and the whole community. The security issues ofhigh-risk jobs have received increasing attention.To analyze the risk factors in the accidents has very important practical and profoundsignificance. This paper provides methods to make statistical analysis of accident riskfactors based on the data of high-risk operations safety accident as the cases fromhigh-risk operation system. Firstly a preliminary timing analysis and disaggregatedstatistical analysis about the accidents is made to gain a global understanding of theaccidents. In the section based on Fault Tree and Bayesian network-based approach forperforming a risk factor analysis, high falling accidents are took as examples. BuildingFault Tree follows these three steps: to determing influencing factors, to determining thesignificance of correlation coefficient and finally to auxiliary expert experience toconstruction. After converting Fault Tree to Bayesian Network, quantitative analysis ismade based on Bayesian Network. It contains node criticality analysis, sensitivity analysisand Bayesian network node diagnostic reasoning to identify the most likely cause chain.In the section analyzed based on Empirical Bayesian theory, reaearch conducts underHuman Factors Analysis and Classification System (HFACS) framework.This articledemonstrated the methods about ultra parameters and posterior probability estimation.Regarding to the Decision Tree, CHIAD algorithm is selected to build it. The resultsobtained in the above analysis can be taken out to get valuable information about theaccident safety management. inadequate of safety supervision and inspection, pooreducatin, Poor working environment are the three imperative aspects to improve to reduceaccidents.