A Computer Aided System for Correlation and Prediction of Phase Equilibrium Data

T.L. Nielsen, Rafiqul Gani

Research output: Contribution to journalConference articleResearchpeer-review


Modelling, simulation and design of chemical processes need accurate and reliable estimation of the properties of the mixtures present in the process. In these situations, the property models are in a 'service' role where they supply the needed properties only when requested. Due to the solution procedures employed in simulation and design, the property models are required to have well-behaved and continuous derivatives for these problems to converge a solution. Furthermore, in process optimisation, the second and third order derivatives of the property models must exist for solution approaches based on mathematical programming. This paper describes the development of a computer aided system for the systematic derivation of appropriate property models to be used in the service role for a specified problem. As a first step, a library of well-known property models ha's been developed and a feature for parameter estimation has been added, so that for any selected model, the necessary model parameters can be estimated, fine-tuned and/or converted from one model to another model. Also, the property derivatives are validated and verified for thermodynamic and mathematical consistency. (C) 2001 Elsevier Science B.V. All rights reserved.
Original languageEnglish
JournalFluid Phase Equilibria
Issue number1-2
Pages (from-to)13-20
Publication statusPublished - 30 Jul 2001
Event14th Symposium on Thermophysical Properties - University of Colorado, Boulder, Colorado, United States
Duration: 20 Jun 200030 Jun 2000
Conference number: 14


Conference14th Symposium on Thermophysical Properties
LocationUniversity of Colorado
CountryUnited States
CityBoulder, Colorado


  • model
  • equation of state
  • activity coefficient

Fingerprint Dive into the research topics of 'A Computer Aided System for Correlation and Prediction of Phase Equilibrium Data'. Together they form a unique fingerprint.

Cite this