Abstract
Most of the quality inspections (including visual inspections) in the production process are mainly conducted by humans (inspectors). There is a problem that judgments vary due to the influence of the physical condition of the inspector and environmental changes, and there is a demand for the development of an analytical method that can be quantitatively and automatically judged. In this paper, we focus on improving the efficiency of quality inspection of cooling motor fans inside electronic equipment, where demand is increasing due to the weight reduction of IT equipment. The subject of the sensory test requires a lot of experience to accurately diagnose the slight difference in the sound emitted by the fan (sound pressure), and the judgment may vary depending on the physical condition of the inspector and changes in the environment. We apply the "Pattern Recognition MTS (Mahalanobis Taguchi System) method" which enables discriminant analysis of inspection events, and approached the optimization of normal/abnormal discrimination accuracy. As a result of the research, it was verified that it is extremely effective as compared with the conventional discriminant analysis method (FFT analysis etc.).
