Research Of Coupling And Prediction Control Of Multiple Quality Characteristics

With the development of science and technology, manufacturing enterprises are facing frequent and unpredictable changes in market conditions. In this case, more and more enterprises realize the importance of improving product quality. To keep the advanced advantage in the fierce market competition, the control to the quality must be emphasized.Due to the complexity of mechanical and electrical products, quality characteristics present with a high level of coupling. Coupling of quality characteristics is a common phenomenon that exists in various stages of product manufacturing, and makes it complex to predict and control quality characteristics during product manufacturing stage. So, it is necessary to do deeply study for the coupling of quality characteristics.First, the paper has made an analysis on the quality characteristics and its classification. Also the decomposition and transfer of quality characteristics during the whole life of a product is analyzed. Then the paper places great emphasis on the study of quality characteristics coupling. The mechanism of coupling is researched and coupling matrix with quality characteristics is defined. Framework of decoupling technology of QCs based on AD is proposed.To evaluate the coupling strength of coupling matrix, a decoupling modeling with Analytic Network Process is established. Fuzzy theory and Clustering Algorithm are introduced to build a more accurate model. Based on the model imported above, solving procedure of decoupling is presented. The illustration of the proposed method is verified by a case study.A coupling and forecasting model was constructed based on general predictive control theory(GPC) for quality characteristics prediction. Least squares identification is introduced to build coupling matrix. Three decoupling algorithms including GPC algorithm with reference observer, algorithm with feed forward decoupling and GPC algorithm with objective function decoupling is presented based on the model to solve the coupling of quality characteristics prediction.At last, an instance of decoupling of quality characteristics prediction in the cutting process is studied. By simulation and analysis of the three decoupling algorithms, a several conclusions are provided:(1) The three decoupling algorithms can reduce the coupling of prediction system to some extent.(2) Algorithm with feed forward decoupling and GPC algorithm with objective function decoupling is more effective than GPC algorithm with reference observer.(3) GPC algorithm with reference observer can combine with other decoupling algorithms, which can improve the prediction accuracy.