a novel approach for phase equilibrium calculations for geological sequestration

 due to excessive amount of anthropogenic co2 emissions, scientists have been trying to develop some reasonable solutions to this problem (greenhouse gases). one of the most reasonable short-term solutions for removing greenhouse gases (mainly co2) is to sequester them in underground formations including seabed as well as geological formations. in this regards, in order to have an exact prediction of the capacity of the mentioned areas to sequester, we must have a clear understanding about co2-brine thermodynamic equilibrium meaning that we must have a precise mathematical model with ability to predict the latter calculations. in this paper, artificial neural network technique as data processing tool was applied to co2-brine tertiary system and consequently was used as our model. in this regards, a feed-forward network architecture was used and later optimized in regards of the number of hidden layers and neurons in each layer by using sensitivity analysis on our target variables. finally, our model was verified by experimental data acquired from the literature. the results show that our model predictions are in close agreement with the experimental data having a regression coefficient – r2 – of about 0.997. also they represent that in some cases, the neural network – based models can be more accurate than equation of state – based ones which makes it reliable for substituting classic methods in the near future.