Slot Die 코팅 공정의 모델 예측 제어를 위한 다중목적 최적화 기법

Abstract

In the field of chemical engineering, several studies proposed to find appropriate operating parameters of the Slot Die coating process, but there is still a problem with product quality because of limitation through trial and error by workers at the manufacturing site. Moreover, PID (Proportional-Integral-Derivative) controller and MPC (Model Predictive Control) are mostly used as process control techniques but they still have some limitations such as nonlinearity, different environmental condition, complexity of response, or multiple objectives of manufacturing processes. In this paper, we propose a novel multiple objective model predictive control technique called MOORA (Multiple Objective Optimization based Regression Analysis) in the Slot Die coating process with multiple-input and multiple-output. In this study, modeling with multiple regression analysis, we derive input values and output target values. Furthermore, in the multiple objective optimization problem that find the best solution, the proposed method has shown better execution time and performance than NSGA-Ⅱ(Non-dominated Sorting Genetic Algorithm).