Optimization of electrospun polyamide 6,6/polyvinyl alcohol nanofibril composite through response surface methodology and neural networks

The present investigation employs artificial neural networks (ANN) and response surface methodology (RSM) to model the diameter of a polyamide-6,6 (PA66)/polyvinyl alcohol (PVA) nanofibril composite. For modelling the average diameter of nanofibers, RSM approach based on four parameters (PA66/PVA (wt.%), high voltage (kV), needle tip to collector distance (cm) and flow rate (mL/h)) in five levels was compared to ANN technique. Central composite design (CCD) was employed in the RSM approach to ascertain the individual and combined effects of the parameters on the average diameter of nanofibril composite. The network topology of 4-10-1 has good predictability for analyzing factors impacting PA66/PVA fiber diameter, as demonstrated by the strong correlation between the experimental and predicted PA66/PVA fiber diameters using an artificial neural network. ANN predicts PA66/PVA nanofibril composite diameter with an average absolute relative error of 1.149%, which is marginally lower than RSM’s result of 1.639%. The high regression coefficient between the variables and the response (R ² = 0.991) shows a good second-order polynomial regression model for evaluating experimental data. The R ² value was 0.989, indicating that the ANN model was good fit with the experimental results. The RSM model determined optimum conditions for electrospinning PA66/PVA nanofiber that can produce fine, reliable, and high-quality nanofibers: PA66/PVA concentration of 50/50 wt.%, high voltage of 19 kV, needle tip to collector distance of 12 cm, and flow rate of 0.1 mL/h. This investigation is novel in that it uses ANN and RSM to compute and model the effects of electrospinning parameters for increasing nanofiber diameter. The investigation will help practitioners, researchers, and engineers create nanofibers using electrospinning machines, which is the scientific application of this investigation.