Phase Equilibrium Modeling in Gas Purification System

Abstract

A thermodynamic model based on activity is proposed to correlate and predict the vapour-liquid equilibria of the aforesaid systems. The activity based models render an insight in to the molecular physics of the system; hence accurate speciation of the equilibriated liquid phase becomes a reality besides its prediction ability of solubility of the acid gases over alkanolamine solutions. The activity based model has been developed using extended Debye-Hückel theory of electrolytic solution with short range, non-electrostatic interactions. The vapor phase non-ideality has been taken care of in terms of fugacity coefficient calculated using Virial Equation of State. The equilibrium constants are taken from literature as functions of temperature only. The neutral and ionic species present in the equilibrated liquid phase have been estimated with zero interaction model and incorporated here. The interaction parameters in the activity models are estimated by minimizing the objective function, which is the summation of relative deviation between the experimental and model predicted CO2 partial pressures over a wide range.
The parameter estimation for the phase equilibrium models have been formulated here as a multivariable optimization (minimization) problem with variable bounds. The MATLAB 7.6 optimization toolbox has been used extensively for the present work. ‘fmincon’ function, which is a constrained optimization function uses quasi-Newton and Sequential Quadratic Programming (SQP) methods, has been used here for minimization of the proposed objective functions with variable bounds for both approximate and rigorous modeling. There remains a necessity of refinement of the developed rigorous thermodynamic model in terms of the accurate speciation, i.e., exact determination of the species concentration in the equilibrated liquid phase and use of better optimization algorithm, may be non-traditional one, which will ensure global minima.