Murata Manufacturing Co., Ltd. Senior Instructor Katsushige Hamaguchi
Plackett-Burman's L12 orthogonal array is an experimental design that takes up 11 factors, but has only 12 experiments. Therefore, it is often used in screening experiments, and statistical model analysis cannot analyze interactions and only evaluates main effects. However, various combinations are performed as experiments, and the coverage rate of 3-factor combinations is 100%, and the coverage rate of 4-factor combinations is about 75%, making it a highly efficient design table in terms of combination evaluation. Using this combination evaluation, we proposed a method to create an experimental design with a high success rate, and many examples have been produced. In this presentation, we will use three representative examples to introduce a method to achieve results with the shortest number of experiments by reflecting technical capabilities in the experimental design. We would like to recommend this to anyone who is worried about the large number of experiments in a 3-level experiment.
