@PHDTHESIS{LPT-diss-2002-01,
AUTHOR = {H. Briesen},
TITLE = {{Adaptive Composition Representation for the Simulation and Optimization of Multicomponent-Mixture Processes}},
YEAR = {2002},
Publisher = { VDI-Verlag, D\"{u}sseldorf},
series = {Fortschritt-Berichte VDI: Reihe 3, Nr. 740},
note = {},
abstract = {The need for more rigorous models for multicomponent-mixture processes as for example refinery processes demands the use of a detailed molecular level composition representation. The size of such models and therefore the computational cost to solve them will in the future increase by orders of magnitude. This work shows the benefits of a composition representation which is automatically adapted to the current accuracy needs. Technically, this adaptivity has been realized by utilizing both mathematical and chemical engineering concepts. By employing a continuous composition representation, a Wavelet-Galerkin discretization, multigrid concepts and a residual based adaptation strategy, it was possible to derive a method which allows to obtain approximations of the solution with a high compositional detail without actually solving the model for this high compositional detail. The proposed method was implemented and tested for a variety of physical systems and process problems},
keywords = {Detailed refinery simulation, adaptive composition, multigrid methods, real-time optimization, petroleum mixtures, pseudocomponents, steady-state optimization, steady-state simulation, dynamic simulation.},
}

Adaptive Composition Representation for the Simulation and Optimization of Multicomponent-Mixture Processes

Abstract:
The need for more rigorous models for multicomponent-mixture processes as for example refinery processes demands the use of a detailed molecular level composition representation. The size of such models and therefore the computational cost to solve them will in the future increase by orders of magnitude. This work shows the benefits of a composition representation which is automatically adapted to the current accuracy needs. Technically, this adaptivity has been realized by utilizing both mathematical and chemical engineering concepts. By employing a continuous composition representation, a Wavelet-Galerkin discretization, multigrid concepts and a residual based adaptation strategy, it was possible to derive a method which allows to obtain approximations of the solution with a high compositional detail without actually solving the model for this high compositional detail. The proposed method was implemented and tested for a variety of physical systems and process problems

Keywords:
Detailed refinery simulation, adaptive composition, multigrid methods, real-time optimization, petroleum mixtures, pseudocomponents, steady-state optimization, steady-state simulation, dynamic simulation.