Abstract
Four developed machine learning algorithms are proposed to prognosticate the CO2solubility in amino acid salt solutions, blended with amine solutions as additives, in broad ranges of temperature and pressure. From literature 375 experimental data points for CO2solubility were collected. The results from the applied algorithms indicated that the CO2solubility is estimated acceptably close to the experimental values. In the best case, the developed network estimates CO2solubility in the stated solutions with an average relative deviation of 6.53 % and a correlation coefficient of 0.9892.

This publication has 44 references indexed in Scilit: