A new convective heat transfer model for disk-shaped particles from flat disks to cylinders derived using direct numerical simulation and genetic algorithms

This study investigates the convective heat transfer characteristics of disk-shaped biomass particles, a critical factor for particle heating and conversion in biomass co-firing boilers that has been underexplored in the literature. Using particle-resolved direct numerical simulation (PR-DNS), we analyze the heat transfer behavior of cold fluid flowing over heated disk particles, focusing on the effects of particle aspect ratio or height-to-diameter ratio (Ar), Reynolds number (Re), and incidence angle (θ) on the average Nusselt number (Nu). The results reveal that heat transfer efficiency is governed by the interplay between fluid dynamics and particle geometry. Increasing Reynolds numbers significantly enhances heat transfer, while incidence angle and aspect ratio further modulate efficiency by altering the particle's projected area. The optimal heat transfer occurs at an incidence angle of θ = 75°, where the average Nu exceeds that at θ = 15° by 13.81%. A novel Nusselt number correlation for disk-shaped particles is developed using genetic algorithms, accurately predicting heat transfer across a wide range of conditions: Re ≤ 2000, 0 < Ar ≤ 1 (i.e., from highly flattened disks to near-spherical cylinders) and 0° ≤ θ ≤ 90°. This model serves as a robust tool for analyzing convective heat transfer of both non-spherical and spherical particles, contributing to improved designs for biomass co-firing systems.