Similarity Measure Supported Fuzzy Failure Mode and Effect Analysis

Abstract

Nowadays in the various engineering fields quality requirements are continuously increasing. There is also a need to develop flexible and highly adaptive systems to meet current requirements. At the same time, it is also essential to predict possible system failures and to address the arising problems appropriately. A widely used approach for predicting and preventing system failures is the Failure Mode and Effect Analysis (FMEA), which accompanies the entire development process and is able to adapt to changes in the system. The conventional method can be improved if fuzzy logic is incorporated into the evaluation. In this way the often arising subjectivity and uncertainty can be handled to ensure a more reliable result. In this paper, the author proposes a Fuzzy-FMEA (F-FMEA) based approach supported by similarity measures, for the system level. In the evaluation fuzzy arithmetic operations are applied to determine the Probability of Failure for the different failure codes. In addition to the single-expert F-FMEA system, the evaluation method that takes into account the opinions of multiple experts is also presented.