Vapour-Liquid Equilibria(VLE) is a topic of importance because of its several areas of applications including designing of process equipments. There are theoretical and thermodynamic models reported in the literature for estimation of VLE, however the accuracy is affected because of more than one component involved in the system. The objective of the present work is to utilise Multilayer Perceptron network for modeling of VLE of ten binary mixtures involving combinations of nine components. Vapour-Liquid Equilibria of binary mixtures reported in the literature have been used in the present work that include acetone-ethyl acetate, acetone-methanol, acetone-hexane, water-1,2 ethandiol, ethanol-acetic acid, ethanol-water, methanol-ethanol, benzene-acetic acid and benzene-water. In present work in addition to these parameters molecular weight of components along with DDB number for individual components are also included. The correlation is to be developed for input parameters, molecular weights of the individual components, equilibrium temperature and system pressure with the output parameters equilibrium liquid and vapour phase compositions. There are 222 data points which are divided in two parts, training and test data sets, containing 201 and 21 points respectively, for developing ANN models 1, 2 & 3 using elite-ANN© . The accuracy of the ANN model 3 developed is within acceptable limits of RMSE of 0. 0894 & 0. 1848 for training & test data respectively. The range of applicability of the developed models for temperature and pressure is 312. 45 to 448. 95 K and 23. 371 to 101. 33 kPa. Thus there is a lot of scope to explore developing ANN models with systems incorporating binary mixtures of several components.

1. Anderson J. A, An Introduction to Neural Networks (Prentice-Hall of India, Pvt Ltd New Delhi), 1999.
2. Rumelhart D E & McClleland Back Propagation Training Algorithm Processing, M. I. T Press, Cambridge Massachusetts, (1986).
3. Fan J Y, Nikolau M & White R E, AIChE, 39 (1) (1993) 82.
4. Hoskins J C, Kaliyur K M & Himmelblau D M, AIChE, 37(1) (1991) 137.
5. Watanabe K, Abe M, Kubota M & Himmelblau D M, AIChE, 35 (11) (1989) 1803.
6. Belsito S & Banerjee S AIChE 44 (12), (1998), 2675.
7. Pandharipande S L & Badhe Y P, Chem Eng World 38(6) (2003) 70.
8. Zamankhan P, Malinen P & Lepomaki H, AIChE, 43(7), (1997) 1684
9. Baratti R, Vacca G & Servida A, Hydrocarbon Processing, (1995) 35.
10. Pandharipande S L, Agarwal R S, Gogte B B & Badhe Y P, Chem Eng World 38 (5) (2003) 78.
11. Pandharipande S L & Badhe Y P, Chem Eng World 38 (8) (2003) 82.
12. Pandharipande S L & Badhe Y P IChE, 45(4) (2003) 256.
13. Pandharipande S L & Mandavgane S A, Indian J Chem Technol, 11 (6) (2004) (820)
14. Pandharipande S L, Bhaise A & Poharkar A, Chem Eng World, 39 (1) (2004) 50.
15. Pandharipande S L & Badhe Y P, J Inst Eng, 84 (3) (2004) 65.
16. Pandharipande S L & Badhe Y P, elite-ANN©, ROC No SW-1471/2004.
17. Hajir Karimi and Fakhri Yousefi Chin. J. Chem. Eng. , 15(5) 765—771 (2007)
18. Maria C lluita, Ion lluita, Faical larachi Chemical Engineering Science Vol 55, Issue 15 2813-2825 (2000)
19. Dortmund Data Bank website: http://www. ddbst. com/en/ddbst/index. php.

Vle Artificial Neural Network Binary System