The Dynamic Quality Control Method Based on Relation Analysis

Abstract

To improve product quality in manufacturing process, a dynamic quality control method based on relation analysis is proposed. With the method, the dynamic regulated principal component analysis is constructed by introducing discount factor to eliminate the autocorrelation via the data, and the limit of multiple control charts is calculated by squared prediction error (SPE) statistics. Then, a dynamic adjusting policy by support vector machine (SVM) is proposed based on control chart pattern recognition. Finally, a case study for applicability is presented to verify the proposed method.