@MISC{LPT-pre-2011-33,
AUTHOR = {N. Kerimoglu and A. Mhamdi and W. Marquardt},
TITLE = {{Model-Based Experimental Analysis of Complex Multiphase Reaction Systems}},
LPTKey = {LPT-pre-2011-33},
}

Model-Based Experimental Analysis of Complex Multiphase Reaction Systems

Abstract:
It is important to establish exact process models for model-based design, optimization and process control. Within these necessary process models, the reaction itself is often the most complex part w.r.t. model identification. This is mainly caused by the high number of (potentially) unknown process steps. In multiphase reactive systems for instance, various kinetic phenomena, including heat and mass transfer, diffusion and multiple reactions are occurring at the same time and may interact in unforeseen ways. For exact process models, however, a detailed identification of all these kinetic phenomena is necessary. In order to identify such mechanistically correct models for complex reaction systems, the so called MEXA (Model-based Experimental Analysis) methodology has been developed at Aachener Verfahrenstechnik – Process Systems Engineering, RWTH Aachen University, over the last 10 years. In this methodology, identification of models is divided into simpler sub-models which substantially reduce the complexity and computational load. For a homogeneous reaction scheme, it can be shown as follows. As a first step flux data are determined from experimental data using mass balances of measured species. Next, together known stoichiometric information, reaction rates are calculated. Finally several competing reaction rate models are tested in order to obtain the best model alternative with parameter values used as initial values for a final simultaneous identification step, where complete model is solved once to obtain statistically sound parameter values. Due to the efficient product separation and catalyst immobilization, multi-phase catalytic systems are favoured in industrial applications. Nevertheless, it is usually difficult to decouple reaction and mass transfer kinetics such that the experimental determination of reaction kinetics is masked by mass transfer effects. A novel incremental identification methodology tailored to homogeneous reaction systems and extended to multi-phase systems in previous works allows for the decoupling of reaction kinetics and mass transfer and thus avoids the uncertainty in reaction kinetics identification by construction. While the method has been demonstrated in a simulation case study before, it is applied in a real experimental study of a multiphase system for the first time in this work.