Stefan Puchner,
"Dynamische Modellbildung und Regelung einer komplexen verfahrenstechnischen Anlage zur Bereitstellung definierter chemischer Stoffgemische"
, 12-2009
Original Titel:
Dynamische Modellbildung und Regelung einer komplexen verfahrenstechnischen Anlage zur Bereitstellung definierter chemischer Stoffgemische
Sprache des Titels:
Deutsch
Englische Kurzfassung:
The present work deals with modeling of a complex dynamic production plant of the Lenzing AG to identify improvement and optimization potentials. For that a physical approach was pursued which was already successfully in use for such plants. The model is used to describe a certain substance concentration composed of 4 substances, which various due to various processes. Two of the concentrations can be directly controlled through dosage, the two others are controlled with substance locks. To achieve the required quality, it is necessary that the substance compositions are in narrow limits and provide constant production conditions. The main problem of the constant concentration, and thus derive the motivation for a simulation model from which improvements, and control concepts will be designed, is the currently missing online measurement of the concentration. At the moment the time delay due to the measuring process is up to 20 minutes, which is not suitable for a conventional feedback system. In the modeling process some basic technical elements like the ideal stirred tank and the ideal tubular reactor have been used for describing tank and pipeline models. For getting realistic simulation models, the chemical mixture and the delay due to lead time have been accounted. Various chemical mixtures on pipe and tube branchings were modeled as ideal, which entails no disadvantages. Another key point of the model are substance locks, which have to eject a certain substance for keeping the production conditions almost constant. The procedural theory is widely known, but one of them has slight imperfections, which could not be resolved in this work, because for that some installation work for getting more measured parameters would have been necessary. The total simulation model includes some optimization parameters that were included in the individual components to customize the model using measured data. For the optimization, all inputs were provided with measured data and