Systematic and Optimization-Based Synthesis and Design of Chemical Processes

摘要

In industrial practice, conceptual process design is typically conducted by repetitive simulation studies, which require detailed design specifications in an early design phase. Guided by heuristics, these iterative solution procedures result in high manual effort and, in addition, no guarantee concerning the quality of the solution can be given. Optimization-based design methods provide a tremendous potential to accelerate and improve conceptual process design. Various authors have therefore suggested the use of surrogate models, which do not require detailed specifications. Others have developed methods for the optimization-based process design by means of superstructure optimization. Marquardt, Kossack and Kraemer (2008) proposed a framework for an optimization-based design of hybrid separation processes, which combines shortcut and rigorous evaluation steps. This framework has been successfully demonstrated for conceptual design of various processes (see, e.g., Kramer, Harwardt, Bronneberg and Marquardt, 2011). In this thesis, the process design framework of Marquardt et al. (2008) is extended towards the optimization-based design of reaction-separation processes. For this purpose, powerful shortcut and rigorous evaluation methods for reactor networks and reaction-separation processes are proposed. It is important to emphasize that all of these methods are developed to be computationally efficient in order to allow an optimization-based design and analysis of large-scale processes. As a consequence, cost-optimal process solutions can be obtained with considerably less effort compared to the use of tedious repetitive simulation studies. It also has to be stressed that the performance of all methods is validated by large-scale industrial case studies. Thus, it is shown that the process design framework can contribute decisively towards the sustainable solution of today's challenging design problems in chemical engineering.