Design and control of coupled reactor/column systems

Coupled reactor/column systems feature a reactor with a distillation column sitting above the reactor to remove low-boiling components or sitting below the reactor to remove heavy components. They offer several advantages over separated units: increased field and reduced energy cost, especially when they are used for reversible reactions. The steady-state design and dynamic control of these coupled units are poorly understood and are rarely described in the literature. The various types of coupled reactor/column systems were studied in this dissertation from the standpoint of both steady-state design and control.; Some interesting and unexpected results were uncovered. (1) In conventional distillation column, there is a unique tradeoff between trays and reflux ratio. In coupled systems, this tradeoff also depends on reactor holdup. Based on the relationship, an optimum condition for each system was obtained. In addition, there is a minimum reactor holdup below which the system is infeasible for a given product purity. (2) The dynamics of coupled systems showed characteristics which were significantly different from the individual units (reactor and column). Openloop stability and multiplicity of a coupled reactor/column system depend on the control structure chosen. We demonstrated that some structures were openloop unstable and have multiple steady-states. Such multiple steady-states need to be avoided. This can be accomplished by selecting a control structure that does not exhibit multiple solutions. (3) The best closedloop control among control structures having been tested was obtained when the temperature in the reactor was not controlled. The best control structure from a binary system worked well with minor modification for more complex systems. (4) Control structure design at the steady-state design stage gives more chance for both optimum steady-state design and good controllability, especially when a system is complex with various interlinked units. Steady-state disturbance sensitivity analysis with rating programs was used to select viable control structures preliminarily in a multicomponent system. (5) The dynamics of an optimum design were compared with the dynamics of several suboptimum designs. These comparisons provided good examples of tradeoffs between steady-state economics and dynamic controllability.