Model fidelity considerations in the dynamic simulation of equilibrium staged separation operations.

Item

Title
Model fidelity considerations in the dynamic simulation of equilibrium staged separation operations.
Identifier
AAI9630465
identifier
9630465
Creator
Huang, Shaoqun.
Contributor
Adviser: Irven H. Rinard
Date
1996
Language
English
Publisher
City University of New York.
Subject
Engineering, Chemical
Abstract
Model fidelity, namely, how well a particular model represents the behavior of an actual system, is an important consideration in dynamic process simulation. As even more sophisticated models are used in order to obtain a better understanding of both the physical phenomena and expected performance of processes, the trade-offs between model complexity and model fidelity must be considered. There is a need for guidelines or criteria to assist systems engineers and dynamic models in the appropriate choice of model as a function of fidelity and complexity.;In this thesis, model fidelity is studied by using a reference model. These other models may be either simpler or more detailed than the reference model. The object is to develop methodology for making informed choices as to which model to use in a given application. Our approach has been to develop model fidelity indices of various types against which model fidelity errors can be correlated. These should be simple to evaluate a priori so that the user can decide which model to use without expending a lot of effort.;Three equilibrium staged separation operations were studied: (1) a single-stage flash drum, (2) an isothermal gas absorber, and (3) a simple multicomponent distillation column. The first two, being relatively simple, were used to develop indices and model fidelity error correlations and guidelines. These were then tested on the more complex distillation column model.;Two aspects of modeling were included in this study. One is the choice of vapor-liquid equilibrium (VLE) model. The other is the choice of what structural details should be included in the model. Two VLE models were compared, namely, ideal and that based on the Wilson activity coefficient model. A number of structural details associated with multi-staged separations were studied.;The results validate the approach, namely, that model fidelity indices can be used to correlate model fidelity errors. Several were developed, the most useful being Index A which allows one to decide what VLE model to use. The results also show that under most circumstances, the vapor holdup can be neglected compared to the liquid holdup. Downcomer dynamics can generally be neglected over the frequency bandwidth of usual engineering interest but flow-over-the-weir dynamics cannot. Most of these models are very stiff and the choice of the appropriate integration method is not always straightforward.
Type
dissertation
Source
PQT Legacy CUNY.xlsx
degree
Ph.D.
Item sets
CUNY Legacy ETDs