Prediction of solubility of ammonia in liquid electrolytes using Least Square Support Vector Machines

Abstract

Abstract Liquid electrolytes or ionic liquids (ILs) are a new class of environmentally friendly solvents which is highly promising for refrigeration and air conditioning applications in chemical industrial processes. In this contribution, Least Square Support Vector Machine (LS-SVM) models have been utilized to predict the solubility of ammonia in ILs as a function of molecular weight (MW), critical temperature (Tc) and critical pressure (Pc) of pure ILs over wide ranges of temperature, pressure, and concentration. To this end, 352 experimental data points were collected from the published papers. Moreover, to verify the accuracy of the proposed models, statistical analyses such as regression coefficient, mean square error (MSE), average absolute deviation (AAD), standard deviation (STD) and root mean square error (RMSE) have been conducted on the calculated values. The results show the excellent performance of LSSVM models to predict the solubility of ammonia in different liquid electrolytes.