A new data envelopment analysis-based model for failure mode and effect analysis with heterogeneous information

•A new DEA-based model is proposed for FMEA with heterogeneous information.•Multiple types of information are are applied for describing different risk factors.•FEC method was employed for clustering heterogeneous risk assessments.•An improved DEA method is introduced to derive risk ranking of failure modes.•The effectiveness of the proposed FMEA is illustrated with an empirical case study. Failure mode and effect analysis (FMEA) is a multi-disciplined and group-based reliability optimization technique extensively used in diverse industries. When applied in real-life situations, however, the traditional FMEA method has been found to be less effective because of its inherent deficiencies. In the risk analysis process, different types of evaluation information may be involved because of heterogeneous features of risk factors. Moreover, the increasing complexity of products and systems necessitates the participation of a large group of experts with various backgrounds and expertise in FMEA. Thus, in the paper, we develop a new risk analysis model based on data envelopment analysis (DEA) for FMEA within the large group environment. Multiple types of information, including crisp values, interval numbers, and uncertain linguistic terms, are applied for describing different risk factors. The individual risk evaluation information is clustered by an extended fuzzy equivalence clustering method. A modified DEA model is proposed for deriving the risk priority ranking of the identified failure modes. Finally, the effectiveness and applicability of the presented new FMEA model are testified by a case study of blood transfusion and a comparison analysis with relevant existing methods.