A computational model incorporating realistic microstructures for predicting effective thermal conductivity of polyimide nanofiber aerogel

Aerogel, characterized by its ultra-low density, high specific surface area, and low thermal conductivity, is a typical porous material. The polyimide (PI) aerogel, incorporating electrospun nanofibers, has garnered significant interest due to its exceptional benefits in flexibility and thermal insulation. Traditional models of thermal property prediction are unsuitable for PI nanofiber aerogel due to its high porosity (>90%) and fiber randomness. This study focuses on constructing a computational micro-mechanical model incorporating stochastic nanofibers to predict the macroscopic thermal properties of polyimide nanofiber aerogel. Fractal analysis of thermal conductivity is conducted, leveraging the self-similarity inherent in porous media. A stochastic fiber generation algorithm is proposed to simulate the random spatial distribution of nanofibers. By controlling the number, size and degree of regularity, a versatile finite element model (FEM) is established with air matrix and fiber elements. Adjustable regularity of fiber distribution and fiber agglomeration effect are achieved to reflect the realistic conditions of nanofiber microstructures. Finite elements of fiber and matrix are coupled via Voronoi tessellation, and the effective thermal conductivity (ETC) is calculated. In comparison with experimental results of self-prepared PI aerogel samples via freeze-drying procedure, the predicted values possess an average error of 1.45% for samples of different densities, demonstrating a satisfactory accuracy of the model.