Sven Kossack

A Framework for the Systematic Design of Hybrid Separation Processes

The design of optimal separation flow sheets for multi-component mixtures is still not a solved problem. This is especially the case when non-ideal or azeotropic mixtures or hybrid separation processes are considered. We review recent developments in this field and present a systematic framework for the design of separation flow sheets. This framework proposes a three-step approach. In the first step different flow sheets are generated. In the second step these alternative flow sheet structures are evaluated with shortcut methods. In the third step a rigorous mixed-integer nonlinear programming (MINLP) optimization of the entire flow sheet is executed to determine the best alternative. Since a number of alternative flow sheets have already been eliminated, only a few optimization runs are necessary in this final step. The whole framework thus allows the systematic generation and evaluation of separation processes and is illustrated with the case study of the separation of ethanol and water.

An efficient solution method for the MINLP optimization of chemical processes

Abstract Process synthesis often involves the solution of large nonlinear discretecontinuous optimization problems, which are usually formulated as mixedinteger nonlinear programming (MINLP) or generalized disjunctive programming (GDP) problems and solved with MINLP solvers. This paper presents an efficient solution method for these problems named successive relaxed MINLP (SR-MINLP), where the model formulations are reformulated to contain only continuous variables. The discrete decisions are relaxed and successively tightened in a sequential solution procedure to facilitate convergence and to obtain local optima of good quality. The solution method is illustrated by a simple numerical example as well as a large and complex example from process synthesis.

Optimal column sequencing for multicomponent mixtures

Abstract The separation of a multicomponent mixture using distillation is usually possible in a large number of different sequences, which will provide the same products but have different energy demand. In this contribution, we provide a systematic method to find the optimal column sequence based on exergy demand. The screening of design alternatives is done within a superstructure framework, which allows for the decomposition of the separation sequences into unique separation tasks. The use of the task concept significantly reduces the computational work. The individual separation tasks are evaluated using shortcut methods. For the application to azeotropic mixtures, the mixture topology is determined and feasibility checks are performed for every split. In this context, azeotropes are treated as pseudo-components.

MEXA goes CAMD — Computer-aided molecular design for physical property model building

Abstract The development of physical property models is an ongoing challenge in chemical engineering. It usually requires both theoretical insight as well as experiments to test and validate the models. Model-based experimental analysis (MEXA) provides a work process for such developments integrating systems tools and experiments. Optimal experimental design is here a key step. In physical property model development, the choice of the optimal test mixture itself is crucial but usually not systematically addressed. For a rational solution to this problem, recent methods for computer-aided molecular design (CAMD) are integrated into the MEXA work process. Thereby, a targeted and efficient approach for physical property model development is achieved. The approach is examplified for the prediction of multicomponent diffusion in liquids.