Comparison between thermodynamic and neural network model in methane hydrate formation process

In present work, an artificial model based on feed forward artificial neural network algorithm was employed to estimate pressure of methane hydrate phase equilibria in systems with salinities. To develop this algorithm, the experimental data for methane hydrate formation condition was collected from different literature. Independent experimental data which were not used in training this algorithm have been employed to examine reliability of developed method. This model was validated using data in various literature. The major characteristic of this ANN model is its ability in prediction of methane hydrate formation pressure in various ranges of temperature and amount of salinity. It is shown a good agreement between experimental data and predicted corresponding values.

کلیدواژه ها:

Methane ، Hydrate ، Salinity ، Artificial neural network