An improved FMEA model considering information quality in a multi-granularity probability linguistic environment

As a significant analytical tool in reliability management, FMEA has been extensively used in various fields. Nevertheless, conventional FMEA has been criticized for some defects. To compensate this situation, this article proposes an improved FMEA method under the environment of probabilistic linguistic terms. The multiformity and indeterminacy of experts' assessment information is depicted by applying probabilistic linguistic term sets, and then evaluation information is fused based on information quality and Dempster-Shafer evidence theory. The different action priority is adopted to determine the sequence of failure modes. Finally, a case study is presented to verify the applicability of the proposed method.