Correlation of binary VLE by GE models: Comparison of Two variants of Pearson’s unconstrained optimization methods
Advancing the chemical engineering fundamentals
Thermodynamics (T2-1P)
Keywords: Pearson’s
Correlation of binary VLE by GE models: Comparison of Two variants of Pearson’s unconstrained optimization methods
A. Ravi Prasad, J. Srinivasa Babji, A. Aravind Kumar and B. U. S. Naidu Babu
Department of Chemical Engineering, Andhra University, Visakhapatnam, India
The binary VLE data are needed in the design of distillation equipment and for testing of liquid mixture models. Most of the liquid mixtures encountered in commercial operations are non-ideal. The non-ideality in a liquid mixture can be concisely represented by the Excess Gibbs Energy . There are several empirical and semi theoretical models proposed in literature. The correlation of P–T–x–y data consists in determining the constants in the GE models by minimizing the objective function . The constants can be obtained by several methods, which may be classified as
Methods based on solution of sets of non – linear equations
Methods based on optimization techniques
The models proposed in literature can be classified as
Single parameter
Two parameter
Multi parameter models
The models can be broadly classified as classical models and models based of local composition concept. Often they are classified as Enthalpic models, entropic models and complete models .
In this paper two variants of Pearson’s unconstrained optimization methods are compared for estimation of parameters in isothermal binary VLE data of the system Hexafluoro benzene (1) / benzene (2) system at 333.15 K using several models and the results of processing the data are presented.
Presented Monday 17, 13:30 to 15:00, in session Thermodynamics (T2-1P).
