Fibrous media are widely utilized for insulating high-temperature industrial equipment due to their excellent high-temperature insulation performance. However, the multi-scale and heterogeneous structure of fibrous media poses considerable challenges for the accurate prediction of their thermal behavior. To overcome these challenges, this study proposes a stochastic algorithm for generating realistic fibrous media models. Compared to computed tomography (CT) techniques, the proposed method significantly reduces modeling costs and improves computational efficiency while preserving essential structural accuracy. Furthermore, Computational Fluid Dynamics (CFD) and the Monte Carlo method are combined to investigate the heat transfer characteristics of fibrous media. The effects of fiber surface emissivity, the degree of fiber arrangement disorder, and the number of fibers in contact with the hot and cold boundaries on the effective thermal conductivity are systematically examined. The results indicate that the effective thermal conductivity decreases monotonically with increasing fiber arrangement disorder. Additionally, both fiber surface emissivity and the number of boundary-contacting fibers show a positive correlation with effective thermal conductivity. The model's accuracy is validated against experimental data, with an average deviation between simulated and measured values below 15 %. This study establishes a theoretical framework for predicting thermal transport behavior in complex fibrous media.

Fibrous Media,3D stochastic geometry,Effective thermal conductivity,Radiation heat transfer