Estimation of Solubility of H۲S in Ionic Liquids by Machine Learning

In various liquid and gas streams, acid gases such as CO۲ and H۲S are hazardous. Consequently, removing these components especially in applications such as chemical production, by absorption method using physical adsorbents, ionic liquids and amines. In the present study, the Adaptive Neuro-Fuzzy Inference System (ANFIS) model to monitor the H۲S and CO۲ solubility in four different ILs was used. The statistical analyses yielded values of ۶.۵۵×۱۰-۶ for MSE, ۵.۲۱×۱۰-۳ for RMSE, ۳.۹۲×۱۰-۳ for MAE, ۳.۲۸×۱۰-۵ for RSE, ۰.۹۹۹ for EVS, and ۰.۹۹۹ for R۲ in the case of the ANFIS model. By comparing the ANFIS and PSO-optimized mathematical models from the previous study from with RK-EoS and ER-EoS, it was concluded that the ANFIS model exhibited superior prediction capabilities for estimating H۲S solubility in ILs. Furthermore, the interaction deriving out of the ۲ models demonstrated acceptable coherence with the data collected from the experiment for CO۲/IL/H۲S ternary mixtures.

کلیدواژه ها:

ANFIS ، ILs ، H۲S Absorption ، RK-EoS ، ER-EoS