Study on the occurrence path and prediction of unsafe behaviours of hazardous chemical storage personnel

Abstract

The storage of hazardous chemicals is a high-risk area prone to accidents, which can result in severe environmental pollution, endanger personnel safety, and cause significant economic losses. By applying an improved HFACS model, tailored for hazardous chemical storage and combined with Bayesian methods, a more effective analysis of unsafe behaviour in personnel can be achieved. First, a model for analysing unsafe behaviour in hazardous chemical storage was developed, and 14 relevant factors were identified. Parameter learning was then conducted to preliminarily determine key nodes. Next, the mutual information matrix was calculated, and sensitivity analysis was performed to update the model. Path analysis was then employed to examine the impact of various factors on three types of unsafe behaviour: skill-based errors, decision errors, and violations. Finally, posterior probabilities were calculated to illustrate the unsafe behaviour analysis method. Violations are significantly influenced by the organisational climate, skill-based errors are more affected by the technical environment, and decision errors are primarily influenced by inadequate supervision and poor operational planning. In hazardous chemical storage working scenarios, attention should be focused on improving the technical environment to prevent skill-based errors, addressing inadequate supervision and operational planning to avoid decision errors, and improving the organisational climate to prevent violations. This study presents a methodology, applicable to the hazardous chemical storage industry, that predicts potential unsafe behaviour based on certain factors and provides insights into the causes of accidents that may not be clearly identified in accident analysis reports.