Developing a Bayesian network model for improving chemical plant workers’ situation awareness

Abstract

Abstract The present study aimed to create a Bayesian network (BN) model to manage and improve workers’ situation awareness. The 12 important variables affecting workers’ situation awareness were determined using the Fuzzy Delphi method and experts’ opinions. The data were collected using a self-administered questionnaire. The BN model was created using the Dempster-Shafer theory. The expectation-maximization algorithm was employed to determine the conditional probability tables. Belief updating was utilized to determine the variables with the strongest effects on situation awareness. Based on performance evaluation criteria of the BN model, the model performance was acceptable. Environmental distraction, safety knowledge, and fatigue were the best predictors of situation awareness. Furthermore, it was found that decreasing environmental distraction and work pressure, and improving safety knowledge were the best intervention strategies to improve workers’ situation awareness. Overall, we can conclude that the BN model is a powerful tool to create a causal model. Moreover, using belief updating as an exclusive characteristic of BN enables managers to select the best intervention strategies. The results of this study provide a basis for managers’ decision-making to improve employee safety performance in the workplaces and the proposed model can potentially be used for employee safety performance.