Fuzzy inference system application for failure analyzing in automobile industry

Nowadays, quality is one of the most important key success factors in the automobile industry. Improving the quality is based on optimizing the most important quality characteristics and usually launched by highly applied techniques such as failure mode and effect analysis (FMEA). According to the literature, however, traditional FMEA suffers from some limitations. Reviewing the literature, on one hand, shows that the fuzzy rule-base system, under the artificial intelligence category, is the most frequently applied method for solving the FMEA problems. On the other hand, the automobile industry, which highly takes advantages of traditional FMEA, has been deprived of benefits of fuzzy rule-based FMEA (fuzzy FMEA). Thus, the purpose of this paper is to apply fuzzy FMEA for quality improvement in the automobile industry.

Firstly, traditional FMEA has been implemented. Then by consulting with a six-member quality assurance team, fuzzy membership functions have been obtained for risk factors, i.e., occurrence (O), severity (S), and detection (D). The experts have also been consulted about constructing the fuzzy rule base. These evaluations have been performed to prioritize the most critical failure modes occurring during production of doors of a compact car, manufactured by a part-producing company in Iran.

Findings indicate that fuzzy FMEA not only solves problems of traditional FMEA, but also is highly in accordance with it, in terms of some priorities. According to results of fuzzy FMEA, failure modes E, pertaining to the sash of the rear right door, and H, related to the sash of the front the left door, have been ranked as the most and the least critical situations, respectively. The prioritized failures could be considered to facilitate future quality optimization.

This research provides quality engineers of the studied company with the chance of ranking their failure modes based on a fuzzy expert system.

This study utilizes the fuzzy logic approach to solve some major limitations of FMEA, an extensively applied method in the automobile industry.