| Summary: | 碩士 === 國立臺灣科技大學 === 纖維及高分子工程技術研究所 === 85 ===
In these years, the application of polymer materials is growing broader in industry, and the commercial products for daily life are also more and popular. Thus the request to the quality of polymer product raises responding. The rheology of polymer melt, viscosity to be an important one, from great effect the quality of products. To improve the quality of products directly, it is necessary to measure and control the viscosity of melt during the process. For now, the methods of changing products quality by controlling the viscosity can be classified into three types according to the character of the measurement of viscosity, say, off-line, on-line and in-line. For the method of in-line, the delay between the time when melt passing by the viscometer and that of output signal being obtained is tiny. This is the only method that allows real-time control and is worth researching.
This paper focuses on the design and analysis of the control strategies for extrusion process of a screw-extruder in a continuous production extruding process. The goal is to control the property (viscosity mainly) of output polymer materials. The viscosity of LDPE is measured by using an in-line viscometer developed in our lab. This in-line viscometer not only benefits of other types lie capillary or slit of defect of flow on the decrease. We consider the extruder as an SISO system, with a process input of screw speed and a process output of the viscosity of polymer melt. Because the relative equation between these two during the process with a good reliability is difficult to retrieve, we describe characters of the dynamic and disturbance process with experimental method.
Accounting for many undetermined factors like uneven feed, the unstability of the machine and screw speed, great variations of the output that cannot be eliminated or repressed by using high-frequency filters and traditional control theory would happen. Therefore, we apply the minimum variance simulation in optimal control theory and limit the screw speed to invoke constrained minimum variance estimation in order to make the result reliable. Then we use the pole placement in the traditional control theory to control the viscosity of melts and compare the result with that of minimum variance control theory for the sake of identifying and improving the feasibility of the system.
|
|---|
