Pore diameter of nanoporous anodic alumina: Experimental study and application of ANFIS and MLR

This paper presents numerical and experimental results to show the effect of nonlinear models has on the prediction of pore diameter of nanoporous anodic alumina (NPAA). For this purpose, a database of pore diameters was collected from the literature to developing two models: (1) an adaptive neuro-fuzzy inference system (ANFIS) model and (2) a multiple linear regression (MLR) model. Inputs for these two models compromised of the concentration, the temperature and the voltage whereas pore diameter in a NPAA membrane was the target output. To illustrate the accuracy of two proposed models, a comparison between outputs obtained from the models, test results and two empirical formulas has been made. The comparison reveals that the models have a good capability of predicting pore diameters with an acceptable error with respect to empirical formulas.