An intelligent scenario for methane hydrate formation at different thermodynamic conditions

A feed forward artificial neural network (ANN) with Levenberg–Marquardt training algorithm was developed to predict the pressure of methane hydrate three phase equilibria. Pressure of methane hydrate formation depends on temperature, type and amount of salinity. These variables are considered as the input of ANN model, and the output of model was compared with experimental data. Statistical parameters r (regression) and MRE (mean relative error) are employed as evidence to value the operation of model. By comparing the results, it can be concluded that ANN predictions is a good approximate for prediction of methane hydrate phase equilibria.

Keywords:

Methane hydrate , Phase equilibria , salinities , Thermodynamic model , Artificial Neural Network