New improved CREAM model for human reliability analysis using a linguistic D number-based hybrid decision making approach

Hua Shi,Jing-Hui Wang,Ling Zhang,Hu-Chen Liu

DOI: https://doi.org/10.1016/j.engappai.2023.105896

2023-01-01

Abstract:Since human errors are the critical factor contributing to accidents in safety-critical industries, human reliability analysis (HRA) becomes significant to improve the reliability of many complex engineering systems. In the real world, the assessment of human errors relies heavily on experts' knowledge and experience. Moreover, there are challenges regarding the human error probability (HEP) estimation because of state evaluation information uncertainty, multiple common performance conditions (CPCs), and the correlations between CPCs. Therefore, this paper aims to propose a new improved cognitive reliability and error analysis method (CREAM) by integrating linguistic D numbers and the decision-making trial and evaluation laboratory-based analytic network process (DANP) method for the quantitative analysis of human errors. First, the linguistic D numbers are utilized to express the uncertain CPC state evaluations and the direct relations between CPCs provided by experts. Second, the DANP method is incorporated into the CREAM to determine the weights of CPCs considering their interdependencies and interactions. Then, a CPC effect index is introduced to quantify the CPCs' overall effect upon performance reliability and obtain the HEP estimation. Finally, the feasibility and practicality of the proposed CREAM model are demonstrated through a practical case of the liquefied petroleum gas (LPG) cargo loading process. The results show that the new improved CREAM can effectively capture the uncertain state evaluation information, better depict the interdependences between CPCs, and obtain more reasonable HEP estimation in the HRA.