(372d) Systematic Screening and Design Evolution for Catalytic Gas Conversion Processes

The past two decades have seen increased research efforts into how optimal process designs can be systematically achieved for a given chemistry using mathematical optimization of process network superstructures (1). Such methods have recently been highlighted as key technologies that promise to deliver major improvements in process efficiencies (2). However, the methods are rarely applied in industry, mainly as a result of their inability to handle practical process constraints, to robustly process complex kinetic models and to support the design-decision making process. We present the development of advanced optimal process synthesis framework that enables the identification of optimal reaction-separation process structures. The framework enables the exploration of design trends and key design features that maximize process performance in terms of profit for catalytic gas phase reaction systems, taking into account various design options for the reactor, the separation system and the energy systems.

The framework introduces the concepts of design screening and evolution and permits the design engineer to explore the system's performance and complexity trade-offs. In the screening stage, the framework allows the screening of vast numbers of potential design candidates to identify optimal and near-optimal conceptual process design candidates. In the evolution stage, these promising design candidates can be explored further in two directions. In terms of design complexity, the relationship between the design features identified in the screening stage and the performance can be established. Furthermore, detailed reaction process models are employed to explore the potential effects associated with non-ideal behavior of the system. Both design stages employ tailored superstructure formulations that are searched for optimal design options using stochastic techniques in the form of Tabu Search.

The framework has been developed by building upon our previous work (1,4,5). We have developed process superstructure representations specifically tailored to heterogeneously catalyzed gas-phase reacting systems. The superstructures account for detailed representations of the reaction system through combinations of generic units that embed options related to mixing, temperature policies, the mass of catalyst present, as well as constraints related to energy management and composition limits. The separation systems are represented in aggregated form. This is achieved through cost models that are developed using separation systems synthesis models and capture the lowest cost separations possible for given stream compositions and flow rates. The model aggregation allows the decoupling of separation synthesis calculations from superstructure optimization and keeps the problem numerically solvable for problems of industrial complexity. Energy targeting methods are employed to ensure designs require minimum amounts of energy.

The process synthesis is performed in two stages to generate maximum design insight and understanding. Screening stage: the performances of conventional designs are obtained as base cases and the performance limit that can be attained for the system with an innovative process design is obtained from superstructure optimization allowing all possible interactions between feed streams and recycle streams with multiple reaction zones each of which can exhibit different mixing, heat management and catalyst mass. Evolution stage: The designs from step 1 are analyzed and key design features are identified for which targeted superstructures are optimized to study the impact of individual design features as well as the effect of non-ideal behavior of the reaction system.

Applications to a number of industrially relevant reaction systems will be presented, including the partial oxidation of ethane to acetic acid (3) and the production of styrene. The results will show how the framework has produced innovative designs with significantly improved performances over their conventional counterparts.

References

1. P Linke, AC Kokossis, AIChE J 49(2003) 1451.

2. C Tsoka, WR Johns, P Linke, A Kokossis. Green Chem 8(2004), 401.

3. D Linke, D Wolf, S Zeiss, U Dingerdissen, M Baerns, J Catal 205(2002) 32.

4. V Ashley, P Linke. Chem Eng Res Des 82(2004), 952.

5. P Linke, AC Kokossis. Comp Chem Eng 27(2003), 733.