Modelling, Design, Operability and Analysis of Reaction-Separation Systems

This thesis presents a model-based Methodology for the design and analysis of chemical processes involving a Reaction-Separation with Recycle (RSR) scheme. At the core of this Methodology, a model-based analysis of firstprinciples models of different complexity identifies the important interactions between design and process variables, so that specified design targets may be achieved.

Based on this, flexible operating policies are developed together with the corresponding control structure in order to maintain the specified design targets. The model-based Methodology divides the analysis into three stages. Stage 1 simplifies the design-control problem such that only the main operations are analysed.

The corresponding model, developed in terms of dimensionless variables, helps to identify the limiting values of a set of lumped (design-process) variables and locates operational windows where process feasibility can be assessed. In Stage 2, more detailed models are developed by delumping the set of lumped variables and relaxing the assumptions that were considered in Stage 1.

The only solutions to be considered are those lying within the bounds (operational window) defined in Stage 1. The objective of Stage 2 is to identify the location of an optimal operation based on a specified design target. The final Stage 3 is employed as a verification step by means of more rigorous models in steady-state and dynamic modes.

The information generated through Stages 1 and 2 serves as the initial estimates for the needed rigorous simulations in Stage 3. The Stage 3 simulations validate the design decisions made from the earlier stages. Through the solution of a number of hierarchical sub-problems, the solution of a complex design-control problem is therefore obtained systematically along with an understanding of the process behaviour.

The model-based Methodology has been applied to a series of RSR systems (case studies) varying in complexity and ranging from first-order reactions with perfect separators to multiple (isothermal and no-isothermal) reaction systems with more elaborate separation section (distillation-based, strippers or flash units).

Through the analysis of theoretical RSR systems, it was possible to demonstrate different process behaviours (input and output multiplicities) under different scenarios of process operation. These issues are of major importance when an operating policy or control structure needs to be developed. For more complex RSR systems, the model-based Methodology has shown its potential to address the important issues related to the integrated design and control problem.

Note that the important design and process variables, once identified, lead to the proposal of a simple control structure that maintains the design specifications and provides the basis for a more thorough analysis.