>From toreh@smekk.hit.no Fri Mar 13 13:34:23 1998

Tore Haug-Warberg, Høgskolen i Telemark:
Implementation of next-generation equilibrium algorithms in MATLAB


                               ABSTRACT

Thermodynamic equilibrium computations have been the subject for almost 40
years. It all started with the development of the RAND and SOLGAS algorithms
initiated by the Apollo-program back in the 1960's. At that time the
chemical equilibrium behaviour of ideal gas mixtures (possibly with some
solid phase precipitates) were needed to give an adequate description of
rocket propulsion systems. Later, in the early 1970's MINISIM was developed
by the late prof. John Erbar to do simple unit operation computations.
During the 1970's other program systems were developed for e.g.
metallurgical problems (ThermoCalc, MTDATA, FACT, etc.) and geochemical
computations (Ge0-Calc). Another direction of evolution was the development
of highly sofisticated software for the solution of entire process
flowsheets (ASPEN, PRCOESS, HYSIM, etc.).

While the first 20-30 years were spent in an effort to make algorithms run
faster and with better convergence properties, the last 10 years have been
spent on user interface and Windows applications. Indeed, surprisingly
little work has been done on the very problem formulation, i.e. how to
generalize the thermodynamic problem description such that more general
constraint situations can be handled effectively. In this context MATLAB has
proved to be a valuable programming language, at least for prototyping=
 purposes.

In this lecture I shall present MATLAB implementations of some thermodynamic
software development milestones. Thereafter a short introduction to the
general thermodynamic equilibrium problem, and how it can (partly) be solved
in a MATLAB, will be discussed. Relevant examples from the CADAS project
will be given. The lecture will conclude with what I find the most severe
limitations of MATLAB and give reasons for why the planned dr.ing. project
"The General Thermodynamic Equilibrium Problem - an Object Oriented
Solution" has to be moved to a C++ platform.

********************************************************************
Tore Haug-Warberg                 Fax:    +47-3557-5210
Department of Technology          Phone:  +47-3557-5233
H=F8gskolen i Telemark              Home:   +47-3551-2190
N-3914 Porsgrunn                  e-mail: Tore.Haug-Warberg@hit.no
Norway
********************************************************************