| Summary: | 博士 === 國立臺灣科技大學 === 纖維及高分子工程研究所 === 87 === In the polymer processes, the end products quality would change in accordance with process variations. These process variations could result from the variations in the raw material (e.g. variations in regrind level, drying conditions, and additive concentration etc.) or in the performances of the molding machine and auxiliary equipment (e.g. signal fluctuation due to imprecise sensors, and inconsistent machine operations etc). An effective closed-loop control to eliminate process variations and to rapidly track the set point of quality factor is of primary importance. But it''s a tough work to achieve the demands of both real-time quality control and no being interfered.
Earlier studies performed quality control by off-line methods. Devices like melt indexer and laboratory capillary rheometer etc were used to detect variations in flow characteristics and processability of the polymer. Further steps of turning operation conditions were needed for control. Since the measurement took place away from the production line, it would take not only more time and manpower but also wasting the material. Latter, in the 80''s, the on-line rheometer was used. Tests were performed by a by-pass instrument in the production line. But, from the view point of real-time quality control, it still existed the significant signal delay resulting from the time required for melt to flow through the transit lines and the gear pump. Recently, the in-line rheometer was proposed, of which the flow characterization model provided is much more relevant to the extrusion process than the off-line data. Besides, it offers immediate cost saving by avoiding delays between product quality assessment.
Several studies have already worked on it. We can draw some conclusions according to these studies: 1. Quality like mechanical, optical, electrical properties and homogeneity etc. can not directly be measured in processing, process variables like die pressure drop, flow rate, torque, viscosity and temperature etc. are generally taken as the quality characteristics. 2. The dynamic of the production line is a rather complex process that has lead to a complicate dynamic model to ensure the polymeric melt with consistent properties. These phenomena have resulted in the needs of complex modeling technique or control strategy.
On the demand of efficient-achieving quality control, in practical applications, we further on the research of a more suitable viscosity model and control system to approximate the complex and nonlinear system. And, in the meanwhile, solve the model-variant viscosity control system in extrusion, such that the quality-consistent end products can be ensured.
The thesis embodies 7 chapters. Chapter 1 covers the introduction of the importance and necessity of viscosity control in the polymer processing. Chapter 2 describes the basic theory of viscosity. Chapter 3 and 4 describes the development of simplified on-line and in-line precision viscometers, respectively, for real time polymer characterization and quality control. In Chapter 5, three different empirical modeling techniques, including continuous, discrete and fuzzy model identification, are proposed to assist analysis of the viscosity control system and controller design. In chapter 6, different control strategies including proportional-integral (PI) controller, the fuzzy gain scheduled PID controller, and fuzzy controller are used to control the melt viscosity by manipulating the extruder screw speed. Remarks on the thesis and goals implemented and further researches are described in Chapter 7.
|
|---|
