Abstract: In order to make traditional FMEA approach more scientific and accurate, a novel FMEA approach for risk evaluation under interval neutrosophic background was proposed, which considers both consensus level and risk preference of decision makers. Firstly, interval neutrosophic numbers were introduced to represent decision makers' evaluation on failure modes and risk factor weights, and a new consensus model was developed to accurately identify and revise non-consensus evaluation information. Secondly, weights of risk factors were determined according to collective evaluation matrix of consensus risk factors combined with evaluation information of risk factor weights and interval neutrosophic entropy measures. Finally, IN-TODIM method was constructed considering irrational psychological behaviors of decision makers to obtain risk ranking of failure modes, and it was also applied to evaluate risks of a steel smelter. The results show that the proposed FMEA approach can iteratively improve non-consensus evaluations that deviate from group opinions and thus obtain more accurate evaluation results.

Key words: failure mode and effect analysis (FMEA), risk evaluation, consensus model, comprehensive weighting method, interval neutrosophic-tomada de decisao interativa e multicritevio (IN-TODIM)