Large-eddy simulations of conjugate heat transfer in boundary layers over laser-scanned ice roughness

Accurate modeling of ice accretion is important for safe and efficient design of aircraft and wind turbine systems. Heat transfer predictions obtained from fluid flow solvers are used as input in ice accretion codes. In glaze ice conditions, freezing rates and resulting ice shapes are highly sensitive to input values of the heat transfer coefficient. Hence, accurate prediction of heat transfer on iced airfoils is crucial for correctly predicting the ice accretion process. In this study, we perform conjugate heat transfer (CHT) simulations using wall-modeled large-eddy simulation (WMLES) over surfaces characterized by ice roughness. The results show that WMLES with CHT accurately captures surface temperature distributions and heat fluxes across a range of roughness geometries. For cases considered, large roughness-to-boundary-layer thickness ratios disrupt outer-layer similarity, leading to substantial errors in estimating equivalent sandgrain roughness when applying traditional empirical models based on surface statistics. The simulations further show that local heat fluxes vary significantly across roughness elements due to low thermal conductivity of the solid; in particular, roughness crests exhibit reduced fluxes in contrast to slopes and valleys. Notably, as roughness height increases, wall heat flux at the crest diminishes, even leading to heat flux reversal in some cases, where thermal energy is transferred from fluid to solid. These effects are not captured in isothermal wall simulations, which overestimate the Stanton number, especially at roughness peaks. By enabling calculation of Stanton number using heat flux distributions, not directly available in experiments, the present simulations augment experimental results and highlight the importance of including solid conduction effects for accurately modeling heat transfer over rough, low-conductivity surfaces such as ice.