A few theoretical studies dealing with hydrate–liquid two-phase equilibrium and three-phase equilibrium in porous media is available in the literature. In this paper, an alternate tool, i.e. the artificial neural network (ANN) technique was employed to estimate the temperature of gas hydrate in porous media. The ANN model was applied to the available 359 and 258 sets of data for methane and ethane respectively comprising temperature pressure and pore size. In order to validate the ANN model, the samples were divided into two groups. One containing 289 and 208 samples of methane and ethane that were employed to train the two networks whereas the remaining 70 and 50 other samples were used for validating the model. Moreover, training of the networks was done applying the standard feed forward back propagation algorithm and several structures were tested to obtain the most suitable network for prediction of the equilibrium. The hydrate phase equilibria for binary mixtures of CH4+water as well as C2H6+water in silica pores were predicted. The predictions of this method were in acceptable agreement with independent experimental data, demonstrating the reliability of ANN tool for estimation of the hydrate phase equilibrium in presence of porous media.

Keywords