Use of the Taguchi Design of Experiments to Optimize the Parameters of Cleaning Machines for Sensitive Plastic Products

Abstract

The Taguchi Design of Experiments (DOE), an optimization technique, was used to establish stabilized operating parameter settings for an ultrasonic cleaning machine for a new and very dirt sensitive plastic product referred to as type A plastic produced in a company in the Philippines. The optimization process is done in actual industrial setting. Taguchi DOE defined quality as minimum variation around the target specification of a product. The matrix used for the project is the La 27 orthogonal array wherein the experiment is comprised of eight runs with seven maximum possible factors at two levels. After each experimental run, the percentage ejection rate was monitored as the measurable characteristic of the finished product. Since rejection rate was the parameter used, the ultimate value closer to zero is better (the smaller the better). Confirmation runs showed that the percentage of rejection rate has improved from 62% to 17% when the established optimized machine parameters were used. Moreover, production yield has increased from 75% to more than 90% for product type A.