chemical engineering research and design 87 (2009) 997–1002 Contents lists available at ScienceDirect Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd Simulation of steam distillation process using neural networks M.T. Vafaei, R. Eslamloueyan * , Sh. Ayatollahi Petroleum and Chemical Engineering Dept., Shiraz University, Zand Street, Shiraz, Fars, Iran abstract Steam distillation process improves oil recovery processes involving steam injection up to 50%. Due to its immense effect on oil recovery, several attempts have been made to simulate this process experimentally and theoretically. Since detailed crude oil data is rarely available, a model should be presented to predict the distillate rate with minimum entry parameters. For this purpose, a Multi-Layer Perceptron (MLP) network is used in this research as a new and effective method to simulate the distillate recoveries of 16 sets of crude oil data obtained from literature. API, viscosity, characterization factor and steam distillation factor are input parameters of the network while distillate yield is the result of the model. Thirteen sets of data were used for training the network and three remaining sets were used to test the model. Comparison between the developed MLP model, Equation of State (EOS)-based method and Holland–Welch correlations indicates that the errors of the MLP model for training and test data sets are significantly lower than that of those methods. Also, the MLP network does not require oil characterization, which is a necessary and rigorous step in EOS and Holland–Welch methods. © 2009 Published by Elsevier B.V. on behalf of The Institution of Chemical Engineers. Keywords: Thermal EOR; Steam distillation; Neural network; MLP 1. Introduction Steam distillation, the process of separating light fractions from the crude oil by forcing steam into a reservoir, is a well- known mechanism in steam injection enhanced oil recovery processes. Steam injection makes lighter fractions of hydro- carbons to be distilled at temperatures lower than their boiling points. Steam and vaporized hydrocarbons will be condensed as they reach cooler regions. Since steam is injected con- tinuously in this process, the condensation and vaporization mechanisms are repeated during the process (Fig. 1). Enhanced oil recovery processes based on steam injection are of the most popular and effective methods used widely in oil recovery industries. Oil displacement in these processes involves simultaneous heat, mass, and fluid transport. Several investigations have been performed to evaluate the contribu- tion of different mechanisms to oil recovery in these methods. Viscosity reduction plays a key role in increasing the oil recov- ery during thermal processes including steamflood technique (Prats, 1986; Farouq Ali, 2003; Nabipour et al., 2007). However, steamflood method differs markedly in performance from ∗ Corresponding author. Tel.: +98 711 2303071; fax: +98 711 6287294. E-mail addresses: eslamlo@shirazu.ac.ir, reslamloueyan@yahoo.com (R. Eslamloueyan). Received 20 September 2008; Received in revised form 23 February 2009; Accepted 25 February 2009 other thermal methods. Several experimental studies have shown that, in addition to viscosity reduction, steam distil- lation plays a significant role in steamflood process. Very low residual oil and much higher recovery of oil after steamflood- ing are because of steam distillation effects (Willman et al., 1961; Wu, 1977). Therefore, steam distillation has been simulated through different methods. Elderly simple mathematical descriptions (Bailey, 1941; Holland and Welch, 1957; Van Winkle, 1967) have inadequate capabilities for predicting crude oil behavior because of their restrictive assumptions. Wu and Elder (1983), proposed correlations to estimate the steam distillation yields with basic oil properties, at saturated steam pressure, with a standard error of 5.6%. Each correlation was dependent on a single parameter and all data points were used in setting up the correlations by regression. Duerksen and Hsueh (1983) showed experimentally that the steam distillation yields at steamflood conditions are sig- nificant even for heavy crude oils. They developed correlations for prediction of steam distillation yield with different crude characteristics and operating conditions. They also found that 0263-8762/$ – see front matter © 2009 Published by Elsevier B.V. on behalf of The Institution of Chemical Engineers. doi:10.1016/j.cherd.2009.02.006