# A surrogate-based inverse design framework for targeted diameter control of electrospun nanofibers

## TL;DR

This paper introduces a data-driven framework to precisely control the diameter of electrospun nanofibers using predictive modeling, reducing the need for trial-and-error experiments.

## Contribution

The novel contribution is a surrogate-based inverse design framework combining XGBoost and PSO for targeted nanofiber diameter control.

## Key findings

- XGBoost achieved high predictive accuracy (R² = 0.890) for nanofiber diameter modeling.
- PSO optimization reached high inverse design accuracy (R² = 0.991, MAE ≈ 1.777 nm).
- Applied voltage and solution concentration were identified as the most influential parameters.

## Abstract

Electrospinning is a high-throughput technique for producing nanofibers. The diameter of such nanofibers governs key properties such as surface area, porosity, and mechanical strength. Precise diameter control is therefore crucial for applications from filtration to tissue engineering, yet optimizing processing conditions for targeted diameter fabrication typically relies on slow, costly trial-and-error experiments. This study presents a data-driven inverse-design framework that replaces traditional trial-and-error optimization with predictive modeling to achieve precise diameter control. Eleven regression models were evaluated on a dataset of 96 poly(vinyl alcohol) (PVA) experiments, with Extreme Gradient Boosting (XGBoost) emerging as the best surrogate (test \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$R^2 = 0.890$$\end{document}). SHAP analysis confirmed applied voltage and solution concentration as the most influential parameters, consistent with physical principles. In the optimization stage, Particle Swarm Optimization (PSO) achieved the highest inverse design accuracy (\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$R^2 = 0.991$$\end{document}, MAE \documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$\approx 1.777\,\textrm{nm}$$\end{document}). This framework enables rapid, efficient design of nanofibers with specified properties and is readily adaptable to other materials and fabrication processes.