A predictive model for electrospun based Polyvinyl alcohol (PVA) nanofibers diameter using an artificial neural...

Area of Science:

• Materials Science
• Nanotechnology
• Computational Modeling

Background:

• Electrospinning is a versatile technique for producing polymer nanofibers.
• Controlling nanofiber diameter is crucial for tailoring material properties.
• Predictive modeling can optimize electrospinning parameters for desired outcomes.

Purpose of the Study:

• To develop a predictive model for polyvinyl alcohol (PVA) nanofiber diameter using Artificial Neural Networks (ANNs).
• To identify key electrospinning parameters influencing PVA nanofiber diameter.
• To determine the optimal ANN architecture for accurate diameter prediction.

Main Methods:

• Artificial Neural Networks (ANNs) were employed to model the electrospinning process.
• Key variables analyzed included electric field, polymer concentration, injection rate, and nozzle-collector distance.
• Various ANN topologies with single and double hidden layers were evaluated to find the best architecture.

Main Results:

• An ANN configuration of 5-9-1 demonstrated strong predictive capabilities for PVA nanofiber diameter.
• The developed model showed high correlation (R² ≈ 0.973) with experimental data.
• A low average absolute error of 0.06 was achieved, indicating model accuracy.

Conclusions:

• ANNs provide a robust tool for predicting PVA nanofiber diameter in electrospinning.
• The model effectively captures the relationship between process parameters and fiber diameter.
• This predictive capability can aid in optimizing electrospinning for specific applications.