PREDICTION OF CO2 SORPTION AT LOW PRESSURE IN POLY IONIC LIQUIDS BASED-ON AMMONIUM USING ADAPTIVE NEURO FUZZY INFERENCE SYSTEM + GROUP CONTRIBUTION METHOD

In this presented work, group contribution method (GC) and adaptive neuro-fuzzy inference system (ANFIS) together have been used to predict the amount of Carbon dioxide sorption in Poly ionic liquids (PILs) based on Ammonium. The modeling is for a dataset containing 7 PILs with 59 data points. 70 of data set has been used for training the networks, 25% of data set for testing and 5% of data set for check data of ANFIS. The model for CO2 prediction in this modelling is an ANFIS model with Hybrid optimization method for minimizing the amount of errors and the cluster center’s range of influence is 0.9. To distinguish Polymeric ionic liquids from each other, the structure of PILs is defined by group contribution method and the number of chemical structures for each of PILs in the form of a matrix is used as an input for the ANFIS model. In addition to the chemical structures, pressure (bar) and temperature (oC) also are used as input parameters to the network.