Quantitative estimation of poly(methyl methacrylate) nano-fiber membrane diameter by artificial neural networks

Abstract

Relationship between the electrospun fiber diameters of poly(methyl methacrylate) (PMMA) nanofibers with process parameters are complex and nonlinear. We used artificial neural networks technique to estimate the electrospun PMMA nanofiber diameter as a function of polymer concentration, nozzle-collector distance, temperature, flow rate, and voltage. The average errors of the predicted fiber diameters for training and testing data were found to be 1.26% and 5.74%, respectively. Process window for optimum nanofiber diameter was generated. The proposed index of relative importance, evaluated in this study, will be a useful guide to quantitatively and qualitatively identify and define the importance of different electrospinning parameters on the fiber diameter. (C) 2015 Elsevier Ltd. All rights reserved.