Design and Control of Heterogeneous Reactive Distillation System

• Main Authors: Yi-Wei Chen, 陳毅偉
• Other Authors: 余政靖
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/34431490031114751263

碩士 === 國立臺灣大學 === 化學工程學研究所 === 92 === Abstract Last decade has seen renewed interest in reactive distillation. Malone and Doherty (2000) summarize recent progresses. Most of the work focuses on steady-state design and analysis for specific chemical systems. In this work, we study a specific class of reactive distillation: heterogeneous reactive distillation (similar to heterogeneous azeotropic distillation). By heterogeneous reactive distillation, we mean two-liquid phase exists in the reflux drum and a decanter is used for liquid-liquid separation. At the conceptual design stage, we look at the appropriateness of using a decanter in producing high purity product. For esterification reactions, heavy esters such as butyl acetate and amyl acetate fall into this category. Next, a systematic procedure is devised for the design of this type of reactive distillation. Results show that feed location is the dominant optimization variable. For example, 33% and 30% energy saving can be achieved for the butyl acetate and amyl acetate, respectively, simply by arranging the feeds optimally. Since the top product composition is determined by the tie-lines, intuitively, we are dealing with a single-input-single-output (bottoms composition) system. Unfortunately, as pointed out by Luyben (2000), we have to control an internal composition (or temperature) in order to achieve material balance for the reaction/separation system. Therefore, we have a 2�e2 control problem. Five control structures are examined and their performance is evaluated. In terms of controlled variables, these five control structures range from composition control, mixed composition and temperature control to temperature control, and, in terms of throughput manipulator, they change from the feed flow to heat input. Disturbances considered are production rate variation, feed ratio imbalance, and catalyst deactivation. Simulation results indicate that reasonable control can be achieved using the candidate control structures (e.g., CS2 and CS5) even under significant catalyst deactivation (~ 50% of original activity).