RSM and ANN approaches for modeling and optimizing of electrospun polyurethane nanofibers morphology

Abstract

This paper focused on using response surface methodology (RSM) and artificial neural network (ANN) to analyze polyurethane (PU) nanofibers morphology synthesized by electrospinning. The process was characterized in detail by using experimental design to determine the parameters that may affect the nanofibers morphology such as polymer concentration, a tip to collector distance and applied voltage. It was concluded that solution concentration plays an important role (relative importance of 79.85 %) against nanofibers diameter and its standard deviation. Based on the results, applied voltage has a different effect on the nanofiber diameter at low and high solution concentrations. Moreover, the tip to collector distance parameter has no significant impact on the average nanofiber diameter. The finest PU nanofiber (201 nm) was obtained from experimental under conditions of: 9 w/v% polymer concentrations, 12 cm tip to collector distance and 16 kV applied voltage. The results show a very good agreement between the experimental and modeled data. It was demonstrated that both models (specially, in case of neural network) are excellent for predicting diameter of PU nanofibers. Furthermore, numerical optimization has been performed by considering desirability function to access the region in design space that introduces minimum average diameter.